Electronic texts and translation series

archive-org-geez: Internet Archive Ge'ez/Ethiopic printed editions

Ge'ez/Ethiopic Christian literature as scanned editions. 210 IA items carry language=gez (all media types); 178 of them are mediatype=texts, of which 143 are openly downloadable and 35 are lending/stream-restricted. Composition of the 178: 76 modern Ethiopian Orthodox Ge'ez liturgical prints (beteyared-book-*, undated, publisher/licence empty), 12 Princeton Theological Seminary Ethiopic manuscripts (ptsem_ethiopian_ms*), 20 Corpus Scriptorum Christianorum Orientalium volumes (CSCO / Peeters / Durbecq / 'Secretaire du CorpusSCO') - only 2 downloadable (actamartyrum0037pere_8 = CSCO Scriptores Aethiopici Ser. Alt. XXVIII; vitaesanctorumin0336gadl), 18 locked behind lending - plus ~49 other printed editions: Budge / Lady Meux 'The Life of Takla Haymanot' 1906 (2 copies + 1 duplicate under india.history.resource.117599), 'The Ethiopic Didascalia' 1834, 'Novum Testamentum ... Aethiopice' 1830, 'Ascensio Isaiae' 1819/1877, 'Das Buch Henoch' 1902, 'Corpus iuris Abessinorum' 1889, 'Gadla Aragawi' 1896, 'Annales Iohannis I' 1903, 'Historia regis Sarsa Dengel' 1907, 'The Book of Thekla' 1901, 'The Ethiopic liturgy' 1915, 'Lexicon linguae aethiopicae' 1865, 'Vitae sanctorum indigenarum' 1904, 'Te'ezaa sanbat' 1902, 'Hermae Pastor' 1860, 'SS. apostolorum Acta et Epistolae Aethiopiae' 1837, 'Vaticinium Malachiae' 1661, 'Homilia Aethiopica' 1660, 'Deo plantante' 1670, 'Le livre d'Henoch' 1892, 'Vida do abba Daniel' 1897. Adding Latin-titled 'aethiopice' editions that lack the gez tag gives a union of 241 texts (151 openly downloadable) - estimate.

Licence: Per-item, not site-wide. 22 of the 178 Ge'ez-tagged texts carry an explicit licence: CC Public Domain Mark 1.0 (18 items, https://creativecommons.org/publicdomain/mark/1.0/ - deed…

Access: No git repository. Enumerate with https://archive.org/services/search/v1/scrape?q=language%3Agez%20AND%20mediatype%3Atexts&fields=identifier,title,creator,date,publisher,language,licenseurl&count=1000 then fetch https://archive.org/download/{id}/{id}_jp2.zip (page images), https://archive.org/download/{id}/{id}_djvu.xml, or https://archive.org/download/{id}/page/n{leaf}_w2000.jpg. Resumable client: 'pip install internetarchive' then 'ia download {id}' (the 'ia' binary is not installed on this machine; the Python package is not installed either). · sample HTTP 200

Not reachable: https://archive.org/about/terms returns HTTP 200 but only a 2,046-byte JavaScript SPA shell; no static Internet Archive terms text is retrievable by curl, so no IA-wide licence wording could be quoted.; https://api.archivelab.org/books/{id}/pages timed out after 30 s (unreachable from this machine) - abandoned in favour of /download/{id}/page/n{leaf}_w{width}.jpg.; https://archive.org/BookReader/BookReaderImages.php?... returns HTTP 404; the working page-image route is /download/{id}/page/n{leaf}_w2000.jpg (200 image/jpeg).; 35 of 178 Ge'ez texts are lending/stream-restricted (18 CSCO volumes among them); {id}_djvu.txt and {id}.pdf return 401/403 and cannot be ingested.; Local OCR is not usable as-is: tesseract 5 lists only chi_sim, eng, osd - no amh/gez/tir traineddata installed.; https://archive.org/services/search/v1/scrape rejects count<100 with RangeException 'count 50 is too small (min count=100)'.

1) Enumerate with the scrape API (count>=100) or advancedsearch; query language:gez AND mediatype:texts, then union with title:aethiopice and subject:'Ethiopic language' because Latin-titled Ge'ez editions are often not tagged gez (12 extra texts). 2) Drop items with access-restricted-item=true (35…

Beta Masaheft (Beta maṣāḥǝft: Manuscripts of Ethiopia and Eritrea — Ethiopic works, manuscripts, TEI)

Ethiopic (Gǝʿǝz) and Amharic Christian literature of Ethiopia/Eritrea, as TEI work records plus catalogued manuscripts. VERIFIED COUNTS (repo tree + API, 2026-11-13): 10,754 work TEI files in git / API total=10,774; 20,537 manuscript records; 18,597 persons; 6,766 places; 923 institutions; 211 narrative units; Clavis Aethiopica index 1,248 works with CPG/BHL/BHO/CC/Syriaca cross-references. Of the 10,754 work files, 4,159 (39%) are IslHornAfr records (suffix 'IHA', Arabic/Islamic, empty editions) — the Christian/Ethiopic corpus is 6,595 works, of which 5,242 are tagged Gǝʿǝz. Genre keywords across those 6,595: ChristianLiterature 5,037, Hagiography 2,128, Miracle 1,368, Poetry/Qene 1,385, Apocrypha 988, Prayers 866, Homily 387, Liturgy 335, Theology 247, Bible 161, CanonLaw 85, Chronography 59. TEXT DEPTH (measured, not estimated): only 373 works carry >=200 words of transcribed edition text (118 of those >=1,000 words); total transcribed Gǝʿǝz text ~1.29M words. 6,461 works have ZERO edition text — these are catalogue records with incipit/explicit/quotation snippets, not editions. Largest editions: Kebra Nagast (LIT1709Kebran, 70,820 words), Jubilees (LIT1697Jubilees, 39,851 words of Vanderkam's edition), 1 Enoch (LIT1340EnochE, 54,519 words), Psalms (LIT2000Mazmur, 41,205 words). Translation coverage is thin: 115 <div type="translation"> blocks totalling ~23.7k words (mostly unlabelled English incipit/explicit glosses); 737 works instead have a BIBLIOGRAPHIC list of published translations.

Licence: CC BY-SA 4.0 per file (10,754/10,754 TEI records carry <licence target="http://creativecommons.org/licenses/by-sa/4.0/">) — CONFLICTS with the site-wide footer, which states CC BY…

Access: git clone https://github.com/BetaMasaheft/Works.git (canonical source TEI, 10,754 files, ~36 MB, branch master) | git clone https://github.com/BetaMasaheft/expanded.git (auto-generated expanded tree: 58,504 files, works 10,744 + manuscripts 20,543 + persons 18,566 + places 6,763 + institutions 926 + narratives 213, ~732 MB, branch main) · sample HTTP 200

Not reachable: curl default User-Agent is reset at TCP level (HTTP 000) by the site WAF — requires a browser-like UA plus cookie ds_verification_token=proof_passed; https://betamasaheft.eu/api/{collection}/list/json?start=N — HTTP 403 Access Denied for every value; only 'perpage' is honoured, so offset pagination is unusable; https://betamasaheft.eu/api/works/list/xml — HTTP 404 (documented in apidoc.html but not implemented); https://betamasaheft.eu/oai?verb=Identify — HTTP 200 but returns the HTML page shell, not OAI-PMH XML; https://betamasaheft.eu/api/dts/document?id=LIT7150Miracle — HTTP 200 with Content-Length 0 (DTS stub); /api/dts/collections and /api/dts/navigation 302 to a self-referential URL; No LICENSE file in any BetaMasaheft repo (GitHub /license API returns 404 for Works, expanded, Manuscripts, Documentation, Schema); DataCite 10.25592/betamasaheft returns rightsList [] — so the licence conflict cannot be resolved from machine-readable metadata

BEST PATH: git clone BetaMasaheft/Works (or the already-current mirror at /srv/christian-corpus/betmas-works, HEAD aebe046 = upstream HEAD as of 2026-11-13T12:01+0100) and parse the TEI locally; do NOT loop the API for 10,774 files. Steps: (1) clone/`git pull` Works; (2) enumerate */*.xml (numbered…

Clavis Clavium (CC) / Clavis Patrum Graecorum (CPG) / Clavis Patrum Copticorum (CPC) / CMCL identifiers

Clavis Patrum Copticorum (CMCL, open, counted by direct harvest 2026-11-13): 1,121 numbered records - cc0001-cc1091 (dense, 3 marked 'vacat': 0058, 0546, 0827), cc2000 (colophon formula), cc5002-cc5097 (29 titulus/'Titulus ad ccNNNN' records). 100 author records in the author index (au@ tokens); the 100 author pages list 589 attributed works, and the anonymous title index lists 379 entries (Bible 79, A 43, B-E 60, F-I 51, J-M 60, N-Q 46, S-Z 40). 1,074 records carry at least one manuscript link (4,580 manuscript-sigla links total). 164 distinct CPG concordances (CG#### - CPG numbers 1830-8800, partial, not the whole CPG) and 80 distinct BHO numbers. 137 distinct 'periodo' literary clusters. TEXT: 0 of the 1,121 records has a digital text - every detail page says 'Testo e traduzione (non disponibile)'; only 7 HTML editions (cc0001, cc0002, cc0065, cc0078, cc0127, cc0415, cc0637) plus ~73 PDFs/plate JPEGs sit under https://www.cmcl.it/~cmcl/testi/ and ~16 book PDFs at https://www.cmcl.it/~cmcl/ . Clavis Clavium (Brepols, paywalled, therefore unverified counts): self-described as 'an integrated reference database and collaborative update platform to open up Patristic, Medieval and Byzantine texts ... to unite the leading claves into one database' (archived Default.aspx, 2022-01-01), i.e. CPG + CPL + CPC + related claves; no entry count could be confirmed because every page is behind the login.

Licence: Split and partly undefined. (1) Clavis Patrum Copticorum on cmcl.it: the entry page calls itself a 'Free service' and the CMCL home page lists it under 'Free online Publications',…

Access: No bulk download, no git repo, no sitemap. Full harvest is ~1,300 GETs against stolet01.cgi (verified: 1,199 responses, no auth, no rate-limiting, no block over ~10 min of sequential requests). Browsable Apache tree: https://www.cmcl.it/~cmcl/ (index of PDFs and testi/ editions). Clavis Clavium has no bulk path at all (login). · sample HTTP 200

Not reachable: https://clavis.brepols.net/clacla/Default.aspx -> 302 /clacla/LoginToolLogin.aspx -> https://apps.brepolis.net/LTool/Entrance.aspx (Brepolis IP/subscription login); Clavis Clavium open-access section: https://clavis.brepols.net/clacla/OA/Browse/Authors.aspx -> 302, https://clavis.brepols.net/clacla/OA/Details.aspx?id=00302FC3BC8C4CCFB00D7DEA3A201284 -> 302; the OA tab has never been anonymously reachable (Wayback CDX shows 302 for every OA capture); https://clavis.brepols.net/clacla/help/manual_en.pdf -> 302 login; not in the Wayback Machine either, so the field schema of Clavis Clavium could not be read; https://www.cmcl.it/~cmcl/agreement_paths_scriptorium_cmcl.pdf (linked from the CMCL home page) -> 404: the only document that might have stated reuse terms is gone; Wikipedia's 'Clavis Clavium ... in open-access' claim (en.wikipedia Clavis Patrum Graecorum) is contradicted by measurement: all CC pages require login; No TEI XML, no JSON, no API, no bulk export, no sitemap, no robots.txt for either resource; clavisclavium.org, clavisclavium.be, clavis.uni-salzburg.at, cpg.bbaw.de: DNS does not resolve (these hosts do not exist)

1) Enumerate authors: GET https://www.cmcl.it/cgi-bin/chiam_clavis.cgi?cl%40cl20=Show -> exactly 100 'name=au@<token> value=Show> <Label>' entries. 2) Per author: ?au@<token>=Show -> each work is a '<p><i>title</i><p><i>title</i> = clavis <cc> (CG####)(BHO####) <br> <b>periodo: </b>...' block. 3) A…

Coptic Scriptorium corpora (public corpora repository)

Coptic (Sahidic and Bohairic) late-antique/medieval Christian literature, 2390 document records in 77 corpora at tag v6.3.1. Breakdown: biblical 2016 documents / 12 corpora - Sahidic OT 911 chapter units (CoptOT, mainly Genesis-Exodus plus Ruth/Jonah), Bohairic OT 507, Sahidica Sahidic NT 259, Bohairic NT 260, plus individual book corpora (Mark 16, 1 Corinthians 16, Bohairic Mark 16, Bohairic 1 Corinthians 16, Jonah 4, Bohairic Jonah 4, Ruth 4, Bohairic Habakkuk 3). Shenoute of Atripe 77 documents / 20 corpora (Acephalous Work 22, Abraham Our Father, I See Your Eagerness, Not Because a Fox Barks, The Lord Thundered, God Who Alone Is True, etc.). Non-biblical, non-Shenoute 297 documents / 45 corpora: Apophthegmata Patrum 126 units; Pistis Sophia 28; Canons of Apa Johannes 14; Theodosius of Alexandria (Encomium on Michael) 9; Martyrdom + Encomium of Victor 10; Coptic magical papyri (Kyprianos) 8; Martyrdom of Helias 6; Besa Letters 5; Literary Fragments 5; Life of Hilaria 5; Pseudo-Theophilus 5; Life of Longinus and Lucius 5; Life of Onnophrius 4; Mercurius 4; documentary papyri 3; Book of Bartholomew 3; Bohairic Life of Shenoute 3; Bohairic Lausiac History 3; Bohairic Life of Isaac 3; Pseudo-Athanasius/Melito 3; Pseudo-Ephrem 3; Pseudo-Celestinus 3; Life of Pisentius 3; plus single/double-document corpora (Life of Marina, Cyrus, Aphou, Phib, Paul of Tamma, John Kalybites, John Chrysostom, Eustathius and Theopiste; Acts of Pilate; Mysteries of John; Dormition of John; Lament of Mary; Gospel of Thomas NHC II; Pachomius Instructions; Proclus Homilies; John of Constantinople; Pseudo-Basil, Pseudo-Cyril, Pseudo-Chrysostom, Pseudo-Flavianus, Pseudo-Timothy). Measured from the files: 1717 CoNLL-U files, 1490 distinct document IDs, 52106 sentences, 1,565,993 tokens; 1458 TEI files (1427 with a real teiHeader, 31 without); 561 TreeTagger .tt files; 1723 relANNIS, 74 PAULA. English translation present for 41,483 / 52,106 sentences (79.6%). Not in scope: Greek or Arabic sources, no Latin, no page images or scans.

Licence: MIXED per document, not a single licence. Counted from meta.json (2390 documents): CC BY-SA 4.0 919; CC-BY-SA (unversioned, all Bohairic NT/OT) 767; CC BY 4.0 (plus 40 with a stra…

Access: git clone --depth 1 --filter=blob:none --sparse https://github.com/CopticScriptorium/corpora.git (then: git sparse-checkout set --cone <corpus-dir>; verified 200/OK, a single corpus checks out at 17 MB). Full tag tarball: https://codeload.github.com/CopticScriptorium/corpora/tar.gz/refs/tags/v6.3.1 (HTTP 200). Versioned archive: https://zenodo.org/records/23676032 (v6.3.1, HTTP 200, single file corpora-v6.3.1.zip, 834,698,296 bytes). · sample HTTP 200

Not reachable: https://copticscriptorium.org/ and https://www.copticscriptorium.org/ - TCP connect timeout to 185.199.108.153:443 (GitHub Pages), repeated twice plus http:// variant; https://pages.github.com/ returns 200 from the same machine, so this is host/IP-specific, not general HTTPS breakage. The site content is available at https://github.com/CopticScriptorium/CopticScriptorium.github.io.; http://corpling.us/ and https://corpling.us/TEI.html - NXDOMAIN (no DNS record); these are the legacy documentation URLs printed in the repo README.; https://corpling.uis.georgetown.edu/coptic-treebank/ - HTTP 404; the Georgetown address cited in coptic-treebank/LICENSE.txt no longer serves the page.; https://data.copticscriptorium.org/urn:cts:copticLit:.../tei/xml|paula/xml|relannis - these look like export endpoints but return 302 to https://github.com/CopticScriptorium/corpora/tree/master/<corpus>/<corpus>_<FORMAT>, i.e. an HTML directory listing, not a file. There is no documented JSON/XML API on the text server.; https://annis.copticscriptorium.org/annis/scriptorium/api/{corpora,annis,version,export,graphml} and /application.wadl - all HTTP 200 but all return the same 1842-byte Vaadin bootstrap HTML (catch-all servlet), so the ANNIS REST API is not usable as documented from here.; http://www.copticscriptorium.org/download/corpora/Mark/coptic_nt_sahidic.html (the canonical Sahidica NT licence page) - unreachable with the host above; licence text was quoted instead from https://web.archive.org/web/2024/http://www.copticscriptorium.org/download/corpora/Mark/coptic_nt_sahidic.html (Wayback, HTTP 200).

Recommended path: pinned sparse git clone, CoNLL-U as the text carrier, meta.json + .tt as the metadata layer. (1) Pin to tag v6.3.1 (commit 6fb51458e7156dd5387c2b2dca207aac1692e4d0, 2026-10-26); releases are also tagged v6.3.0/v6.2.0/v6.1.0/v6.0.0. The README still advertises v6.2.0 on Zenodo (con…

Coptic Old Testament project (CoptOT) - Digital Edition and Translation of the Coptic-Sahidic Old Testament, Goettingen Academy of Sciences and Humanities

Sahidic (plus Bohairic witnesses) Old Testament. Edition text, docID 1620025 'SahidicBible': 82 book units / 1,393 chapters / 37,539 verse units (29,579 OT verses, 7,960 NT verses), 11.99 MB TEI per full pull - 55 OT units including deuterocanon (I-IV Maccabees, Tobit, Judith, Wisdom, Sirach, Psalms of Solomon, Odes, Prayer of Manasses, Baruch, Ep. Jeremiah, Susanna, Bel, Prayer of Azariah) plus the 27 NT books. Manuscript register LCBM: 278 Sahidic biblical manuscripts (XML feed), of which 166 have published diplomatic transcriptions (verified live: 112 return 'No Transcription Available'). Web catalogue: 33 OT book pages, 21 with a critical-edition (collation) route. Apparatus: per-verse segment readings via variant/apparatus/get (Gen.1.1 returns 4 readings with sigla). No modern-language translations are exposed (dataset G is planned/print).

Licence: CC BY-SA 4.0 (stated inside every TEI export: teiHeader/publicationStmt/availability; the project DMP additionally lists the edition text as CC-BY and the manuscript images as per…

Access: https://github.com/CopticScriptorium/corpora (git clone; sahidic.ot/ = CoptOT Basetext, CoNLL-U/TT/PAULA zips; Zenodo DOI 10.5281/zenodo.17917497) - CoptOT itself ships no archive; its API is one call per work · sample HTTP 200

Not reachable: coptot.uni-goettingen.de: NXDOMAIN - no DNS record from this machine (curl exit 6); the live host is coptot.manuscriptroom.com (HTTP 200, 68,790 bytes); api.github.com/orgs/coptot and /orgs/CopticOldTestament: HTTP 404 - no GitHub org, so there is no git mirror of CoptOT's own data; https://data.copticscriptorium.org/texts/old-testament/<doc>/tei: HTTP 500 (text-server TEI endpoint broken); the HTML view /analytic returns 200; https://copticscriptorium.org/: connection timeout (curl exit 28) from this machine; data.copticscriptorium.org is up; /community/vmr/api/ is Disallow-ed in robots.txt for crawlers (Crawl-delay 10); bulk use is welcome only by prior contact

1) Edition text (whole Sahidic Bible): GET transcript/get/?docID=1620025&format=tei&pageID=ALL -> single 11,988,389-byte TEI (HTTP 200, ~19 s). No auth, no cookies. 2) This one request is the fastest complete text: <div type="book" n="Gen">/<div type="chapter" n="Gen.1">/<ab n="Gen.1.1">, one <ab>…

first1kgreek: First Thousand Years of Greek (OpenGreekAndLatin/First1KGreek)

Whole corpus (counted from the upstream git tree at commit ed91552, 2026-10-27): 309 author-groups (CTS textgroups, each with data/<group>/__cts__.xml), 1,103 distinct works, 1,204 TEI editions under data/<group>/<work>/*.xml; catalog.json declares 275,030 nodes, 25,580,179 Greek words, 5,245,040 Latin words. It is the whole first-millennium Greek library, not a Christian-only source: pagan medicine (Galen 103 editions, Hippocrates 53), rhetoric (Libanius 103), philosophy (Aristotle 45), science, scholia all dominate by file count. CHRISTIAN SUBSET, classified by me at author level from data/<group>/__cts__.xml groupname (48 of 309 groups): 249 works, 277 editions, ~8,700,232 words (sum of catalog.json wordcount; estimate, not an independent count). Breakdown: Septuagint tlg0527 = 55 works / 62 editions covering the full LXX from Genesis to Bel et Draco (Pentateuch, Psalms, Odes, Proverbs, Wisdom, Sirach, Psalms of Solomon, all twelve prophets, Isaiah, Jeremiah, Baruch, Lamentations, Epistle of Jeremiah, Ezekiel, Susanna/Daniel/Bel in both Old Greek and Theodotion); Catenae in Novum Testamentum tlg4102 = 33 editions; Origen tlg2042 = 47 works / 1.23M words; Eusebius of Caesarea tlg2018 = 13 works; John Philoponus tlg4015 = 8 works / 1.54M words; Cyril of Alexandria tlg4090 = Contra Julianum (424k words); Epiphanius tlg2021 = 3 works (374k); Chronicon Paschale tlg2371 = 257k; Methodius tlg2959 = 11 works; Theodoret tlg4089; Socrates Scholasticus tlg2057; Sozomen tlg2048; Philostorgius tlg2058; Gelasius tlg2768; Gregory of Nazianzus tlg2022 = 7 works; Gregory II of Cyprus tlg9006; Athanasius tlg2035 = 6 works; Marcellus of Ancyra tlg2041; Clement of Alexandria tlg0555 = 5 works; Clement of Rome tlg1271 = 9 works; Ignatius of Antioch tlg1443 = 2; Polycarp tlg1622; Martyrium Polycarpi tlg1484; Hermas tlg1419 (Shepherd, 31.5k words); Barnabas tlg1216; Diognetus-adjacent Pinytus ogl0001; Justin Martyr tlg0645 + Pseudo-Justin tlg0646; Irenaeus tlg1447; Hippolytus tlg2115; Athenagoras tlg1205; Tatian tlg1766; Theophilus of Antioch tlg1725; Anonymus contra Cataphrygas tlg1131; Seniores apud Irenaeum tlg1665 / Seniores Alexandrini tlg1664; apocrypha and hagiographica: Acta Joannis tlg0317, Acta Thomae tlg2038, Acta Philippi tlg2948, Acta Barnabae tlg2949, Martyrium Perpetuae tlg2016, Testamentum Abrahae tlg1701, Liber Enoch tlg1463, Vitae Prophetarum tlg1750, Marcus Diaconus (Vita Porphyrii) tlg2806, John Zonaras tlg3135 (3 works, 12th c., outside the stated 1st-millennium window). CRITICAL LIMIT: the New Testament is NOT here. Group tlg0031 ("New Testament") contains exactly one work, tlg002, with two files: tlg0031.tlg002.INTF-grc1.xml (INTF 2017) and tlg0031.tlg002.1st1K-cop1.xml (Coptic Scriptorium, Sahidic). The Greek file holds a single textpart running to verse 45 — i.e. Mark 1:1-45 only, 827 words per catalog.json (a complete Mark is ~11k words). Any NT pipeline must go elsewhere (e.g. coptic/betmas or a dedicated NT source). Translations are thin: corpus-wide only 73 non-Greek editions (45 eng, 21 lat, 7 ger); within the Christian subset only 11 groups have any translation and only 7 files are English (Isaiah in 3 editions, Acta Joannis, Eusebius, Acta Thomae, Acta Philippi). The rest are Greek-only.

Licence: CC BY-SA 4.0 (Creative Commons Attribution-ShareAlike 4.0 International). Repo-root license.md is the full CC BY-SA 4.0 legalcode; .zenodo.json declares "license": "CC-BY-SA-4.0";…

Access: git clone --depth 1 https://github.com/OpenGreekAndLatin/First1KGreek.git (verified 2026-11-13: 49.7 s, 984 MB on disk, 311 dirs under data/); single archive https://codeload.github.com/OpenGreekAndLatin/First1KGreek/tar.gz/refs/heads/master (210,718,121 bytes) or .../zip/refs/heads/master (213,098,376 bytes) · sample HTTP 200

Not reachable: https://scaife-cts.perseus.org/api/cts?request=GetCapabilities -> curl exit 35, OpenSSL 'tlsv1 unrecognized name', HTTP 000: the CTS API host no longer has a certificate for that name; https://scaife.perseus.org/api/cts?request=GetCapabilities and ...GetPassage&urn=... -> HTTP 200 but 3,544 bytes of text/html (Scaife Viewer app shell), not CTS XML: no CTS endpoint on that host; https://atlas.perseus.org/api/cts -> curl exit 6, could not resolve host; https://ci.perseids.org/api/hook/v2.0/repo/OpenGreekAndLatin/First1KGreek -> HTTP 404 (the badge API the README advertises is gone; badge images in README.md are dead); https://raw.githubusercontent.com/OpenGreekAndLatin/First1KGreek/master/data/tlg1443/tlg004/tlg1443.tlg004.1st1K-grc1.xml -> HTTP 404, because data/__cts__.xml advertises a work (Ignatius tlg004) that has no file in the tree; filename discovery must come from the git tree; GitHub unauthenticated REST API is limited to 60 requests/hour (measured remaining 39 after probing); a recursive tree is 1 request, so use it once and cache, and use git clone rather than the contents API for bulk

1) Clone shallow, do not scrape raw.githubusercontent per file (1,204 files x individual requests is slow and rate-limited, though raw has no auth and returned 200). `git clone --depth 1` gives the whole corpus in ~50 s at ~1 GB. 2) Enumerate ONLY data/<group>/<work>/*.xml excluding __cts__.xml — t…

HMML / EMML Ethiopian manuscript microfilm and vHMML Reading Room (with HMML Authority File)

EMML (Ethiopian Manuscript Microfilm Library / Ethiopian Manuscript Microfilm Library, Addis Ababa) in vHMML Reading Room: 7,865 digitized manuscript objects, project numbers EMML 1–9232 (7,861 distinct numbers; ~15% of the number range has no record — destroyed or never filmed). They carry 12,745 catalogued text items (parts[].contents[]) described by 217 distinct uniform titles, spread over 736 repositories in Ethiopia and Eritrea. Genres: liturgical books 2,782, hagiographies 1,029, psalters 897, hymns 760, bibles 691, prayers 552, rituals 477, poems 435, gospels 429, missals 391, homiliaries 296, antiphonaries 273 (genre tags over-count, one object can carry several). Languages of the texts: Geʻez 11,901 items, Amharic 926. Dated parts: 6,400, median 1822 CE (p10 1700, p90 1925). Extents recorded for 5,213 objects totalling 633,839 leaves. 1,825 of the 7,865 EMML objects have no content items at all (filmed, not yet described). HMML's own claim on https://hmml.org/collections/eastern-christian: "EMML microfilm collection of 9,000 Ethiopian manuscripts—the largest in the world—photographed throughout Ethiopia during the 1970 and 1980s" — 9,000 filmed vs 7,865 with Reading Room records. IMPORTANT: this is manuscript cataloguing, not editions. There is no transcription, no translation and no critical text for any EMML item; the only full text anywhere is OCR of the printed catalogue (Macomber I–III, Getatchew Haile IV–X = EMML 1–5000) on archive.org. Adjacent Ethiopian collections in the same Reading Room and same schema, useful for the same pipeline: EMIP 4,487 objects, IES 1,867, EMDA 472, UNESCO 364, GG 219. Whole Reading Room holds 264,038 objects across all countries.

Licence: Metadata: Creative Commons Attribution 4.0 International (CC BY 4.0) — HMML terms page: "The manuscript metadata in HMML Reading Room is available for reuse under a Creative Commo…

Access: https://www.vhmml.org/datasets/vhmml_rr_fulldata.zip (48.4 MB zip -> 723 MB JSON, 264,038 objects, Last-Modified 2026-11-09, site text says 'created nightly'); lighter: https://www.vhmml.org/datasets/vhmml_rr_listingdata.zip (3.4 MB -> 55 MB JSON, id/country/city/repository/shelfmark/HMML project number/permalink only); CSV equivalent: https://www.vhmml.org/datasets/vhmml_rr_listingdata_csv.zip. Archive.org EMML catalogues: https://archive.org/details/EMMLCatalog01 … EMMLCatalog10 (metadata API https://archive.org/metadata/EMMLCatalog03, OCR text https://archive.org/download/EMMLCatalog03/EMML%20Catalog%2003%20OCR_djvu.txt). · sample HTTP 200

Not reachable: POST https://www.vhmml.org/readingRoom/search -> 403 ("Turnstile validation failed"); the Reading Room and Data Portal search/export endpoints cannot be driven programmatically.; GET https://www.vhmml.org/dataPortal/export/full and /dataPortal/export/listing -> 400 with any parameter set tried (?keywords=EMML, ?searchType=&keywords=EMML, ?hmmlProjectNumber=EMML%20770); they need a live Turnstile-backed search session, so per-subset exports are unavailable — use the nightly zip instead.; GET https://www.vhmml.org/image/manifest/10218 -> 403 "You do not have permission to access the requested manifest" (accessRestriction "Registered"); same for 10540. Manifest access is object-specific, not global.; GET https://www.vhmml.org/sitemap.xml -> 404, but an official index exists at https://www.vhmml.org/sitemaps/sitemap_index.xml (8 sitemaps, 6 of them readingroom_*.xml at 50,000 URLs each, matching the 264,038-object dataset).; https://hmml.org/collections/manuscripts/ethiopia/ and https://hmml.org/projects/ethiopia/ -> 404 (guessed paths); the real pages are /collections/eastern-christian and /about/global-operations/Ethiopia.; HMML API work: POST https://api.haf.vhmml.org/search -> 500 when `typeList` is omitted (needs typeList ["work"|"person"|"place"|…]); POST /search/related/manuscript/{id} -> 500. No vHMML-side OAI-PMH: /oai, /oai/oai.aspx, /api, /api/ all 404.

1) Pull https://www.vhmml.org/datasets/vhmml_rr_fulldata.zip nightly-or-weekly; it is a single JSON array (one object per line, 264,038 entries) — stream it line-wise instead of json.load to keep memory flat. 2) Filter EMML with hmmlProjectNumber startswith "EMML" (7,865), not collection=="EMML" (7…

Bibliotheca Augustana (Ulrich Harsch, Hochschule Augsburg) — century-by-century Latin library

A one-scholar HTML library organised by period (latinitas romana, mediaevalis, nova) and then by century, verified: saecp05.html (5th century AD, HTTP 200) lists Physiologus Latinus, Historia Augusta, Symphosius, Pelagius, Augustinus, Sulpicius Severus, Claudian, Paulinus of Milan, Paulinus of Pella, Servius, the Haeretici dossier, Patrick, Sedulius, Martianus Capella, Prosper, Salvian, Sidonius. The Augustine page (aug_intr.html, HTTP 200) lists his whole output and links the works actually transcribed: De civitate Dei (aug_cd00.html, HTTP 200) opens with the statement that its text is 'Corpus Christianorum. Series Latina XLVII/XIV, 1 et 2 ed. B. Dombart/A. Kalb, Turnhout 1955', and Confessiones (aug_c000.html, HTTP 200) opens by naming the Maurist/Migne PL 32 text alongside Knöll 1896, Verheijen 1981 and O'Donnell 1992. NOT contained: no apparatus, no line or column numbers, no statement on the text page itself of which of the named editions was followed, no search, no API; the coverage is one man's selection across all of Latin literature, so the Christian share is a thin slice of each century, and there is no medieval depth beyond a few named authors.

Licence: No reuse terms anywhere: a_impressum.html (HTTP 200) gives only the author's postal address and a German liability disclaimer for external links. The underlying texts are PD editi…

Access: https://www.hs-augsburg.de/~harsch/augustana.html · sample HTTP 200

Its value is narrow but real: it is the only source here that gives a clean, readable, complete-looking HTML text of De civitate Dei and the Confessiones with no login, no OCR noise and stable URLs, which makes it a good typological control when OCR from IA or HathiTrust looks doubtful. Because nob…

Brepolis (Brepols) — Library of Latin Texts (LLT), Corpus Christianorum Series Latina (CCSL), Continuatio Mediaevalis, Archive of Celtic-Latin Literature, Monumenta Germaniae Historica online, and Corpus Christianorum in Translation (CCT)

Stated in Brepols' own coverage document (llt_coverage.pdf, HTTP 200, 150,391 bytes) and on the series pages (both HTTP 200): LLT spans classical antiquity, 'Aetas patrum I' (to 500), 'Aetas patrum II' (501–735, to Bede), 'Medii aeui scriptores' (736–1500) and 'Recentior latinitas' (1501–1965), with thematic blocks for the Stuttgart Vulgate, the Latin Old Testament Pseudepigrapha, and Latin translations of the ecumenical councils. Named patristic content: Tertullian, Minucius Felix, Cyprian, Novatian, Victorinus of Pettau, Ambrose, Jerome, Augustine (with the Sermones ad populum re-edited), Hilary, Cassian, Leo the Great, Marius Victorinus, Orosius, Paulinus of Nola (Dolveck), Prudentius, Quodvultdeus, Salvian, Sedulius, Sulpicius Severus, Victor of Vita, Tyconius, plus Aldhelm, Arator, Bede, the Regula Benedicti and Regula Magistri, Boethius, Caesarius, Cassiodorus, Ennodius, Fulgentius, Gregory the Great, Ildefonsus, Isidore. Medieval coverage is explicitly partial but includes Anselm, Bernard, Peter Lombard's Sentences, Aquinas, Albert, Bonaventure, Eckhart, Wycliffe, Gerson, the Legenda aurea, the Ordines Romani (Andrieu), the Spicilegium Friburgense liturgical volumes, and the Innocent III registers. CCSL itself is defined on its series page as critical editions of all Latin Christian texts from Tertullian to Bede (1st–8th c.), with full apparatus and manuscript introduction. NOT open: nothing here is free — brepolsonline.net returns 403 to this host, and access is by institutional subscription, which makes the best critical texts of the Latin corpus unavailable to an open pipeline.

Licence: Proprietary, subscription only; no reuse of the texts or their apparatus is permitted. The Corpus Christianorum in Translation series is a separate paid product and, critically, a…

Access: https://about.brepolis.net/library-of-latin-texts/ · sample HTTP 200

For the gap register this source has two distinct uses. (1) It is the authority to cite for 'the critical text exists' — a work edited in CCSL or CCCM has a modern apparatus even when no open text can be shown, and the register should say so rather than leave etext empty. (2) Its CCT list is a trap…

Corpus Corporum (University of Zurich, mlat.uzh.ch) — the largest openly readable Latin text corpus, with Migne's Patrologia Latina as its core

Measured from its own XML navigation API (php_modules/navigate.php?load=/ returned HTTP 200, 19,693 bytes): 30 top-level corpora totalling 226,311,326 words. Patrologia Latina (corpus idno 38): 5,277 texts, 5,204 works, 1,528 authors, 85,539,587 words — Augustine 5.15 M, pseudo-Augustine 0.89 M, Rabanus Maurus 2.81 M, Jerome 2.31 M, Rupert of Deutz 2.21 M, Innocent III 1.70 M, Gregory I 1.56 M, Bede 1.56 M, Peter Lombard 1.17 M, Ambrose 1.10 M, Bernard of Clairvaux 0.79 M, Anselm 0.36 M, Tertullian 0.34 M. Acta Sanctorum: 66 texts, 52,680,193 words. Monumenta Germaniae Historica: 406 texts / 405 works / 138 authors / 6,643,810 words — a small fraction of the ~97 MGH volumes (see dMGH). 'Scriptores Ecclesiastici' (non-Migne editions, including the Collectio Avellana and Eugippius): 272 texts / 5,845,390 words. Also Thomas Aquinas 77 texts / 8.10 M words, Poetica 681 / 2.82 M, Mirabile 407 / 6.31 M, Latinitas antiqua 197 / 5.26 M, Latinità Italiana del Medioevo 348 / 4.37 M, Neolatinitas 30 / 7.75 M. NOT contained: no Corpus Christianorum (CCSL/CCCM) critical texts, no apparatus of any kind, no Vetus Latina, no manuscripts, no Greek beyond a 775-text miscellany, and no author-by-author coverage of medieval Latin after 1216 — that period appears only as three miscellaneous collections (Latinità Italiana del Medioevo 195 authors / 4.37 M words, Croatiae auctores Latini 331 / 4.53 M, Mirabile 407 / 6.31 M), the Thomas Aquinas corpus, and humanist Neo-Latin. PL coverage is bounded by Migne's own print run, quoted on the series title page: 'ab aevo apostolico ad aetatem Innocentii III (ann. 1216) pro latinis'. No list of PL volume numbers is published, so completeness against the 221-volume print cannot be verified; 14 of the 1,528 author records carry 0 words and 318 carry under 1,000, leaving roughly 1,200 substantive authors.

Licence: No licence or terms page is published. /about/ is an Apache directory index ('Index of /about'), and every candidate terms URL (/terms.html, /about.html, /impressum.html, /help/)…

Access: https://mlat.uzh.ch/ · sample HTTP 200

Access pattern, all verified: navigate.php?load=/<corpus>/<author>/<work> returns XML with word counts at every level (load=/38 returned 477 KB listing 1,528 authors); the per-text TEI is at php_modules/download.php?type=file-xml&idno=<text-idno> (idno 9744 = Abbo of Fleury, Epistolae: 200, 124,882…

Documenta Catholica Omnia (documentacatholicaomnia.eu) — work-level PDFs of Migne and other Latin corpora

A one-maintainer site that mirrors Migne and many other Latin/Greek corpora as per-work PDFs, which makes it the only source here with work-level granularity for Migne rather than volume granularity. Structure verified: index -> 25_10_MPL.html (Migne PL section) -> 20_40_Ecclesiae_Patres_Latini.html (Latin Fathers) -> author page, e.g. 20_40_0354-0430-_Augustinus,_Sanctus.html, which alone lists 573 links, one per work; each link opens a wrapper page whose HREF points at a real PDF (e.g. 02m/0354-0430,_Augustinus,_Ad_Donatistas_Post_Collationem,_MLT.pdf, HTTP 200, application/pdf, 3,818,574 bytes, 20 pages). NOT contained: no OCR text layer behind the PDFs (they are image PDFs of Migne columns), no search, no API, no metadata beyond the filename, no stable identifiers (the filename is the identifier, and it is not a URL-safe string), and its holding of any given work is not verifiable in advance — you must walk the index. The Migne section also carries the Greek series (MPG) mixed into the same site.

Licence: Migne's text is public domain; the site itself publishes only a generic 'Copyright. Terms and Conditions' page (HTTP 200) that explains what copyright is rather than granting a li…

Access: http://www.documentacatholicaomnia.eu/ · sample HTTP 200

Two transport problems are verified and worth recording rather than working around: the HTTPS certificate fails verification ('unable to get local issuer certificate'), so only http:// works, and the site is a single point of failure maintained by an individual. Its unique value is the work-level s…

HathiTrust Digital Library — the Patrologia Latina serial record

Verified through the public Bibliographic API (no key needed, HTTP 200): the PL serial record (OCLC 13948857, record 103439803) carries 260 items, from University of Virginia (157) and University of California (103), and all 260 return rightsCode 'pd' with usRightsString 'Full view' — i.e. the entire PL run is publicly readable as page images plus HathiTrust's own OCR, an independent scan set from Internet Archive's and often a better one. NOT contained: nothing beyond what the partner libraries scanned (this record is PL; MGH, CSEL and CCSL are separate records and were not verified here); no text-level segmentation; no API to pull OCR at scale — the Data API needs OAuth credentials, and the full-text search API needs an institution.

Licence: Public domain, stated per item by the API as rightsCode 'pd' / 'Full view' for all 260 PL items. That is a US determination for a pre-1929 print; reuse of the OCR layer additional…

Access: https://catalog.hathitrust.org/api/volumes/brief/oclc/13948857.json · sample HTTP 200

Important access asymmetry, verified: the API answers 200 from this host, but every HathiTrust HTML page — catalogue search, babel full-text search, even https://www.hathitrust.org/ — returns 403 (5.5–5.9 KB error page) with or without a browser user agent, so the collection is machine-discoverable…

Internet Archive — Corpus Scriptorum Ecclesiasticorum Latinorum (CSEL, Vienna)

The CSEL — the Vienna critical series of Latin patristic editions, still the standard apparatus for many authors before CCSL — is on IA under the stem corpusscriptorum<4-digit-volume>unse (plus Google-scan variants corpusscriptorum<digits>...goog). The phrase search returns 178 items and resolves to 69 volumes of the first 89 (missing 34, 41, 42, 44, 47, 59, 61, 68, 70, 71, 74, 76, 77, 79, 80, 83–88). Verified item: corpusscriptorum0006unse_b8r8 (CSEL 6, Vindobonae 1866) has a _djvu.txt of 2,110,162 bytes (HTTP 200) and a Text PDF. NOT contained: the post-1929 volumes (roughly the second half of the series, including many of the modern critical editions), and anything published only in print after that; no volume-level index, so you must know the CSEL volume number to find a text; no text-level segmentation, so a single CSEL volume's several works arrive as one OCR stream with the apparatus interleaved.

Licence: Public domain for the pre-1929 volumes (CSEL began in 1866); items carry no licenseurl/rights field, so the publication date governs. Note that CSEL editions remain in copyright i…

Access: https://archive.org/advancedsearch.php?q=%22corpus+scriptorum+ecclesiasticorum+latinorum%22&fl%5B%5D=identifier&fl%5B%5D=volume&rows=300&output=json · sample HTTP 200

Why this matters for the Latin slice: CSEL is where the good text of Tertullian, Cyprian, Ambrose, Hilary, Jerome's letters, Faustus of Riez, Eugippius, the Collectio Avellana and dozens of minor authors lives, and Corpus Corporum only carries a 5.85 M-word 'Scriptores Ecclesiastici' selection of i…

Internet Archive — Mansi, Sacrorum Conciliorum Nova et Amplissima Collectio (and the Acta Sanctorum volumes on the same platform)

Mansi's 53-volume collection of councils (Florence 1759–1798) is the only broadly available Latin conciliar corpus that reaches past the ninth century, which is exactly where dMGH's Concilia stop. The phrase search returns 33 items, with two parallel upload families (sacrorumconcilio0001phil..0039ioan and sacrorumconcilio00NNjoan) plus Google-scan bub_gb_ copies and a Mansi33 item; volume numbers recovered from metadata include 1, 3, 20, 21, 28, 29, 33, 34, 39. Verified item: sacrorumconcilio0001phil (tomus I, 1759) has a _djvu.txt fetched at HTTP 200, 4,045,719 bytes. NOT contained: no complete, verified run — the 33 hits overlap in duplicate uploads and do not add up to 53 volumes; the later continuations (Mansi's supplements, Petit/Poujoulat, the Coleti reprint) are separately and patchily uploaded; the metadata language field is wrong on at least one item (sacrorumconcilio0001phil says 'eng' for a Latin volume), so do not trust it; and the text is the 18th-century print with its own apparatus, not a modern critical edition of any council.

Licence: Public domain (18th-century print). No licence field on the items; the usual IA caveat applies — nothing is asserted, so the print's PD status governs.

Access: https://archive.org/advancedsearch.php?q=%22sacrorum+conciliorum+nova+et+amplissima+collectio%22&fl%5B%5D=identifier&fl%5B%5D=volume&rows=30&output=json · sample HTTP 200

Two different needs are served here and should not be conflated: Mansi gives the conciliar acta (canons, letters, subscriptions) in Latin for councils from Nicaea to Trent, i.e. the raw material of the canon-law dossier; Acta Sanctorum (66 volumes, 52.68 M words) is available both here as scans and…

Internet Archive — scanned volumes of Migne's Patrologia Latina (Series Latina)

The PL volumes sit under a single predictable identifier stem, patrologiaecursu<4-digit-volume>mign (e.g. patrologiaecursu0102mign = PL 102). That stem, queried through the advancedsearch API (HTTP 200, numFound 266), yields 176 items covering 168 distinct volumes of the 221-volume print run; every one of the 176 has a _djvu.txt, 176 have a Text PDF and a JP2 zip, and the OCR text totals 751,630,978 bytes. A looser title search ('patrologiae cursus completus', numFound 1,542, of which 646 Latin) finds 209 distinct volume numbers, so other uploaders fill part of the gap: 12 volumes were absent from both queries (8, 17, 18, 23, 31, 38, 40, 41, 66, 73, 77, 185). Verified item: patrologiaecursu0102mign metadata HTTP 200, djvu.txt HTTP 200 at 3,447,804 bytes / 514,337 words. NOT contained: no PL volumes 8, 17, 18, 23, 31, 38, 40, 41, 66, 73, 77, 185 as of this check; no text-level or work-level segmentation (each file is a whole volume, with Migne's prefaces, dissertations, notes and indices interleaved); no critical apparatus beyond what Migne printed; no clean text — the OCR is a search aid, not an edition.

Licence: Public domain: Migne's volumes were printed 1844–1865, and the items themselves declare nothing — metadata for patrologiaecursu0102mign has licenseurl null and rights null, collec…

Access: https://archive.org/advancedsearch.php?q=identifier%3Apatrologiaecursu*+AND+mediatype%3Atexts&fl%5B%5D=identifier&rows=1000&output=json · sample HTTP 200

Verified fetch: https://archive.org/download/patrologiaecursu0102mign/patrologiaecursu0102mign_djvu.txt returns 200, text/plain, 3,447,804 bytes. OCR quality is uneven and must be measured before it is trusted: the running text of PL 102 (Irenaeus in Latin) comes out readable ('DISSERTATIONES PRAEV…

The Latin Library (thelatinlibrary.com) — plain-HTML Latin texts, with dedicated Christian and Medieval sections

Verified section pages: christian.html (HTTP 200, 3,950 bytes) has 57 unique links, 55 of them authors or works — Abelard, Alcuin, Ambrose, Anselm, Aquinas, Arnobius, Arnulf of Lisieux, Augustine, Bede, Benedict, Berengar, Bernard of Clairvaux, Bernard of Cluny, Biblia Sacra, Bonaventure, Carmen de Martyrio, Cassiodorus, the Creeds, Commodianus, the Decretum Gelasianum, Dies Irae, Egeria, Ennodius, Eucherius, Eugippius, Gregory the Great, Gregory VII, Gregory IX, Hugh of St Victor, Hydatius, hymns, Jacobus de Voragine, Innocent III, Isidore, Jerome, Junillus, Lactantius, Leo the Great, the Liber Pontificalis, Novatian, the Passio Perpetuae, Paulinus of Nola, Prosper, Prudentius, Sedulius, Sulpicius Severus, Tertullian, Thomas à Kempis, Vincent of Lérins, the Vita Sancti Columbae. medieval.html (HTTP 200, 5,719 bytes) adds 81 more (83 unique links), from Abbo of Fleury and the Annales regni Francorum through Einhard, Fredegar, the Gesta Francorum, the Historia Brittonum, Carmina Burana, Dante and Johannes de Plano Carpini. NOT contained: no critical editions, no apparatus, no manuscript information, no completeness guarantee — the about.html page admits the texts come from 'different sources... originally scanned... from texts in the Public Domain', some downloaded from dead sites, many 'submitted by contributors', and cred.html frequently says 'from an unidentified edition'; some files are excerpts only. There is no API, no TEI, no citation scheme beyond bare page URLs.

Licence: No licence is asserted: about.html and cred.html (both HTTP 200) explain provenance and credit contributors but grant no reuse terms. Individual texts are PD editions; the HTML ma…

Access: https://www.thelatinlibrary.com/ · sample HTTP 200

Use it as a fast sanity text for a passage, never as the basis of an edition or of a word count. Verified sample: https://www.thelatinlibrary.com/august.html returns 200 (4,846 bytes). The section lists are the useful artefact: they are a rough checklist of which Latin Christian authors somebody th…

Perseus Digital Library — canonical-latinLit (GitHub) and the Scaife Viewer

Inventory measured from the GitHub git-tree API (HTTP 200, truncated=false): 988 files, of which 523 are edition or translation XML and 409 are __cts__.xml metadata, across 60 textgroups — a classical canon (phi*) plus a Stoic fragment corpus (stoa*). The Christian subset is 8 textgroups and 60 files: Tertullian 30 works (stoa0275, opp-lat1), Boethius 12 files including English translations (stoa0058), Prudentius 9 (stoa0238), Augustine 3 (stoa0040.stoa011 = 'Epistulae. Selections.', lat1/lat2/eng1, ed. J. H. Baxter), Sidonius Apollinaris 3 (stoa0261), Jerome 1 (stoa0162.stoa004), Minucius Felix 1 (stoa0203), Pseudo-Tertullian 1 (stoa0276). NOT contained: essentially the whole patristic and medieval Latin corpus — no Ambrose, Gregory, Bede, Isidore, Bernard, Anselm, Aquinas, no councils, no canon law, no hagiography, no liturgical texts; and the Augustine file is a selection of letters, not a corpus. Individual legacy files are TEI P4 with no <availability> element (verified on stoa0040.stoa011.perseus-lat1.xml, 347,175 bytes), so per-file licensing rests on the repository licence alone.

Licence: CC BY-SA 4.0 for the repository (GitHub API license field spdx_id CC-BY-SA-4.0); attribution to the Perseus Digital Library, Tufts University is required by the project's licensin…

Access: https://github.com/PerseusDL/canonical-latinLit · sample HTTP 200

Verified access: raw XML at https://raw.githubusercontent.com/PerseusDL/canonical-latinLit/master/data/stoa0040/stoa011/stoa0040.stoa011.perseus-lat1.xml (HTTP 200, 347,175 bytes); a Tertullian work resolves in Scaife (https://scaife.perseus.org/library/urn:cts:latinLit:stoa0275.stoa001.opp-lat1/js…

Aethiopica: International Journal of Ethiopian and Eritrean Studies (open-access PDFs of Gǝʿǝz text editions, Hiob Ludolf Centre, Universität Hamburg)

Gǝʿǝz (classical Ethiopic) Christian literature as scholarly text editions with facing/parallel translation, published 1998-2025. VERIFIED COUNTS (live, 2026-11-14): 28 volumes (vol. 1/1998 - vol. 28/2025); OAI-PMH catalogue = 1,074 records (articles, miscellanea, review articles, reviews, dissertation abstracts, obituaries). Of those, 41 abstracts explicitly state that the contribution is or contains an edition and/or translation ('edition', 'edited', 'critical text', 'text and translation', 'unpublished'), and 187 records mention edition/ hagiography/ apocrypha/ homily/ hymn/ manuscript keywords. Not every one is a Gǝʿǝz text edition: many are studies, catalogues, or Amharic/Arabic material, so treat 41 as an upper bound. I verified 13 as genuine Gǝʿǝz text editions by downloading the PDF and counting Gǝʿǝz codepoints (list below, with per-article page counts and measured Ge'ez word counts). Subjects covered by these 13: apocryphal acts (Acts of Mark), patristic homilies in Ethiopic version (Cyril of Jerusalem, Yaʿqob of Serugh, Ephrem on Peter), a genuine Cyprianic letter surviving only in Ethiopic (Ep. 70), Christian apocalyptic (Ezra/Salathiel apocalypse; Conversation of Christ at Jericho), a late-Aksumite/Zagwe commentary fragment on Matthew, two malkəʾ hymns (Hripsime; Malkəʾa ʿālam, CAe 6911), a 16th-c. Wǝddase Mäsqäl, a 1596 colophon (Vat. et. 44), one post-classical Gädl (1894, flagged). Editorial standards are high (six-manuscript editions, apparatus, EMML/ms shelfmarks); translations are English for the 2017+ volumes' Ethiopian-studies authors (Haile, Butts, Dickinson, Hensley, Elagina) but Italian, French or German in the older European-authored pieces. TEXT EXTRACTABILITY is volume-dependent and is the single most important fact about this source: volumes 1-19 (1998-2016) embed Ge'ez in legacy 'Ludolfus*' CID fonts with no ToUnicode map, so pdftotext returns ZERO Gǝʿǝz codepoints (measured: art/650 vol.1, art/532 vol.2, art/192 vol.10, art/92 vol.12, art/36 vol.13, art/414 vol.14, art/858 vol.17, art/927 vol.19 all = 0 U+1200-U+137F); volume 20 (2017) onward uses 'EthiopicUnicode' with ToUnicode and extracts cleanly (measured Gǝʿǝz word counts: art/1873 7,088; art/1166 6,927; art/2308 1,673; art/962 1,472; art/2309 1,003; art/1874 676; art/2090 601; art/2283 391).

Licence: not stated

Access: No archive, tarball, git repo or FTP. Bulk = (1) OAI-PMH sweep for the catalogue, then (2) one HEAD/GET per PDF: the galley id is embedded in the article page as article/view/{articleId}/{galleyId}; the file itself is https://journals.sub.uni-hamburg.de/aethiopica/article/download/{articleId}/{galleyId}. Measured PDF sizes 0.32-21.8 MB (median ~1.7 MB); 13 listed articles total 49.4 MB (49,444,105 B). Politeness: robots.txt at https://journals.sub.uni-hamburg.de/robots.txt is exactly two lines, 'User-agent: *' and 'Disallow: /cache/' (HTTP 200), i.e. article and PDF routes are crawlable; no 429/403 seen in ~45 requests at ~3 req/s. · sample HTTP 200

Not reachable: TITUS (http://titus.uni-frankfurt.de/texte/texte2.htm, HTTP 200): its whole 'Ethiopic' section is a single pointer - 'Related content: Thesaurus Linguae Aethiopicae project (Mainz)' - with zero Ge'ez texts hosted. Not a text source.; Thesaurus Linguae Aethiopicae (http://www.uni-mainz.de/Organisationen/TLA/info/ueber.html) -> HTTP 404; the Ge'ez corpus TITUS points to is dead. No successor found.; https://ebible.org/find/details.php?id=gez -> HTTP 200 but the body is 'gez not found / This entry may have been removed or renamed': eBible.org has no Ge'ez (gez) translation, only Amharic/Tigrinya. Also checked https://ebible.org/find/ (200, 769,455 B) - no gez entry.; https://api.getbible.net/v2/translations.json -> HTTP 200, 117 translations, NONE in Ge'ez/Ethiopic (filtered on 'geez'/'ethiop'). Holy-bible APIs do not carry a Gǝʿǝz text.; https://bolls.life/get-translations/ -> HTTP 404 (the documented endpoint moved); abandoned.; https://geez.org/ -> curl exit 28, connection timed out after 25 s (000) - the historical Ge'ez portal does not answer from this machine. Canonical alternative: none; use Aethiopica/Beta Masaheft.; https://tewahedo.org/ -> HTTP 522 (Cloudflare origin timeout), no content.; https://www.eotcmk.org/ -> curl exit 6, could not resolve host (NXDOMAIN); the EOTC North America diocese domain is gone.; https://olac.ldc.upenn.edu/language/gez -> connection timed out (000) after 25 s; the OLAC Ge'ez language page cannot be used as a discovery layer from here.; https://polyglotta.ub.uio.no/ -> curl exit 6, NXDOMAIN (Bibliotheca Polyglotta, which would have carried parallel Ge'ez/Latin texts, no longer resolves).; EAP (Endangered Archives Programme, https://eap.bl.uk/ 200, 362,621 B): its search is POST-only - GET https://eap.bl.uk/search?q=ethiopia returns HTTP 405 - so the Ethiopian project pages cannot be enumerated by URL pattern; images are also partner-rights. Not pursued for a text dossier.; https://github.com/cpradmin/ethiopian-bible (CC BY-NC-SA 4.0, README: 'Manuscript source: Beta Masaheft TEI XML ... 84 books, 1,819 chapters, 37,136 verses, 538,363 words') - REJECTED as unverifiable: Beta Masaheft holds >=200 words of edition text for only 373 works (per the register's measured count), so a verse-continuous 81-book Gǝʿǝz Bible cannot derive from it; the repo also documents an 'IAM automated Ge'ez-to-Latin engine'. Provenance of the Ge'ez column is unexplained, so treat as a derivative of unknown origin, not as a text source. Same caution for biniama/ethiopic-bible-data (Apache-2.0 on the repo, no text provenance) and beldados/EthiopicBibleAPI (MIT; wraps the same unknown text).; https://www.ethiopianorthodox.org/ (HTTP 200, the EOTC's own site) was checked and NOT selected as a text source: its 'Liturgy / Anaphoras, In English and Geez languages' index (https://www.ethiopianorthodox.org/churchmusic/liturgyinenglish/geeze.html, HTTP 200, 12,445 B) contains 0 Ge'ez codepoints - the Ge'ez offerings are MP3 recordings plus ../../biography/englishethiopianliturgy.pdf (English only), and the Amharic archive (amharic/archivee/**/*.pdf, amharic/yezemametsheft/**, biography/books/fethanegest.pdf) is Amharic, unsigned as to licence, and legacy-font. Adjacent, restricted, no bulk path.; https://journals.sub.uni-hamburg.de/aethiopica/api/v1 and .../api/ -> 404 (this OJS install exposes no REST API; OAI-PMH is the only machine route).; https://journals.sub.uni-hamburg.de/aethiopica/about/editorialPolicies -> 404, so the journal's copyright policy had to be read off the individual article pages instead.

1) Build the catalogue from OAI-PMH, not from HTML: GET https://journals.sub.uni-hamburg.de/aethiopica/oai?verb=ListRecords&metadataPrefix=oai_dc then follow resumptionToken to the end (11 requests, completeListSize=1074, ~59 s). Parse dc:title / dc:creator / dc:description / dc:date / dc:subject;…

Patrologia Orientalis: the Arabic-text fascicles (Internet Archive item 'Patrologia-Orientalis'; 64 fascicles, each as an image PDF plus a jp2 page-image bundle and djvu.xml, with EPUB for 60 and a searchable '_text.pdf' for 55)

64 fascicles of the Patrologia Orientalis (Paris/Turnhout, 1904-2010) as individual PDFs, each with EPUB, jp2 page images, djvu.xml and an OCR text file. 13 of the 64 print Christian Arabic texts (Arabic text plus the series' French or English translation), all verified from the fascicle's own title page. The Arabic block: (1) PO 1,2 (n.2) + 1,4 (n.4), History of the Patriarchs of Alexandria, Arabic text ed./trans. B. Evetts 1904; (2) PO 1,3 (n.3), Le Synaxaire arabe jacobite, ed./trans. R. Basset 1904; (3) PO 3,2 (n.12), Severus ibn al-Muqaffa', Refutation d'Eutychius (Le livre des conciles), ed./trans. P. Chebli 1905; (4) PO 4,3 (n.17) + 7,2 (n.32) + 13,4 (n.65), Histoire nestorienne inedite (Chronique de Seert), ed./trans. A. Scher 1907ff; (5) PO 5,4 (n.24) + 7,4 (n.34) + 8,3 (n.38), Agapius of Menbidj, Kitab al-'Unvan, ed./trans. A. Vasiliev 1909-1912; (6) PO 10,2 (n.47), Les menologes des evangeliaires coptes-arabes, ed./trans. F. Nau 1913; (7) PO 10,4 (n.49), Martyrologes et menologes orientaux XVI-XVIII (al-Biruni, al-Maqrizi, Ibn al-Qula'i), textes arabes ed./trans. R. Griveau; (8) PO 18,5 (n.90), Yahya ibn Sa'id al-Antaki, Histoire I, ed./trans. Kratchkovsky-Vasiliev; (9) PO 26,5 (n.129), Voyage du patriarche Macaire d'Antioche (Paul of Aleppo), texte arabe + French, ed. B. Radu (1949, repr. 1990); (10) PO 40,4 (n.185), Une correspondance islamo-chretienne entre Ibn al-Munaggim, Hunayn ibn Ishaq et Qusta ibn Luqa, ed. K. Samir 1981; (11) PO 42,3 (n.192), al-Safi ibn al-'Assal, Brefs chapitres sur la Trinite et l'Incarnation, ed. K. Samir 1985; (12) PO 47,4 (n.212), Histoire de Yahya ibn Sa'id d'Antioche (suite), Arabic critical text Kratchkovsky + French Micheau/Troupeau 1997; (13) PO 49,4 (n.220), The Arabic Life of Severus of Antioch, ed./trans. Y. N. Youssef 2004. Text volume measured: 64 '_djvu.txt' files, 36,287,466 bytes total, of which all 64 contain ZERO Arabic codepoints (U+0600-U+06FF) - the Arabic pages were OCR'd with a Latin/Greek engine and come out as mojibake. A sibling item, 'patrologia-orientalis-29', carries whole volumes 20, 24 and 26-49 (separate PDF + _djvu.txt per volume) and there the Arabic OCR is present but bad: PO 20 9.39% Arabic codepoints, PO 40 4.31%, PO 47 6.36%, PO 49 8.49%, with only 8.9-15.2% of Arabic tokens being common words. Nothing here overlaps First1KGreek/Perseus/Coptic Scriptorium/CoptOT/Beta Masaheft/PEMM/HMML/IA Migne/DigiVatLib; the CSCO Arabic editions (Eutychius, Agapius's CSCO text) are a different edition from these PO fascicles.

Licence: not stated

Access: Whole item, no auth: 119 PDFs (1.66 GB = 64 image-only fascicle PDFs + 55 searchable '_text.pdf'), 60 EPUB (47.8 MB), 64 '_djvu.txt' (36.3 MB), 64 '<fascicle>_jp2.zip' page-image bundles (7.11 GB), 64 '<fascicle>_djvu.xml' (word boxes) and 64 hOCR files; 10.82 GB for every file in the item. · sample HTTP 200

Not reachable: https://www.qdl.qa/en -> HTTP 403 (HTTP/2 403, 5,353-byte block page) on GET with a browser User-Agent, and 403 again on a HEAD retry with a different UA; the Qatar Digital Library cannot be checked from this host. For the same British Library Arabic Christian manuscripts the canonical alternative is the BL's own catalogue, not testable from here today.; https://sinaipalimpsests.org/ -> curl exit 60, SSL: no alternative certificate subject name matches the target host (DNS resolves to 164.67.48.160); over http:// it answers 200 but only serves a static brochure (16,480 bytes). The images live behind a login at https://sinai.library.ucla.edu/ which returns 200 with 'Welcome, please login' and /terms-of-use/please-read - i.e. a login wall, not an open corpus. For Sinai Arabic manuscripts the open route remains the LOC 'Manuscripts in St Catherine's Monastery' collection already registered as greek-ms.; https://biblia-arabica.de/ -> NXDOMAIN (no DNS record from this machine); the Biblia Arabica project site could not be tested, so no claim is made about it.; https://dlme.stanford.edu/ (Digital Library of the Middle East) -> NXDOMAIN from this machine.; https://arabiccollectionsonline.org/ -> NXDOMAIN; the live host is https://aco.dlib.nyu.edu/ (200). Its home page runs window.location.replace('https://aco.dlib.nyu.edu/upgrade.html') and that target returns 404.; https://ebible.org/Scriptures/arb_usfm.zip -> HTTP 404, and the https://ebible.org/Scriptures/ index lists no 'arb*' files, so no Arabic USFM text is reachable under that code.; https://archive.org/download/Patrologia-Orientalis/page/n40.jpg -> HTTP 200 but the body is 'Error serving request: Image error: Metadata error: Couldn't find image stack'; https://archive.org/download/Patrologia-Orientalis/<fascicle>.pdf/page/n40.jpg -> HTTP 404.; Arabic Collections Online has no full-text or OCR layer: its own search page states 'Searches you perform on this site will query only the descriptive information about each book, both in English and Arabic, but not the full texts of the books.'

1) Enumerate: GET https://archive.org/metadata/Patrologia-Orientalis -> JSON 'files'; keep names ending '.pdf' but drop '*_text.pdf' (searchable derivatives, same bad OCR), and keep '*_djvu.txt' for the translation column only. File names carry commas, apostrophes, plus signs, parentheses and space…

armenian-georgian: Armenian and Georgian Christian sources — TITUS Old Armenian and Old Georgian corpora as the text core, plus the Index of Digitized Armenian Manuscripts, the Matenadaran digital library and Grand Catalogue, the Sinai Palimpsests Project of St Catherine's, Armenian Wikisource and the IAE 'House of Holy Manuscripts' collection

Two traditions, three layers. TEXTS (TITUS, the only large free corpus of both languages): OLD ARMENIAN - 16 open text units covering the Zohrab Bible (1805 Venice) NT (Mt-Apoc, chapter pages armnt001-029+ verified) and OT (armat001, 63,634 B), the Old Armenian Lectionary (Lectionarium Hierosolymitanum, 76,202 B), Euthalius Opera (1,031,116 B in one page), Agathangelos Patmowt'iwn hayoc' (25,343 B), Koriwn Vark' Mastoc'i (19,202 B), Eznik Kolbac'i De Deo / Vasn Astucoy (54,290 B), Movses Xorenac'i Patmowt'iwn Hayoc' (58,748 B), Movses Kalankatowac'i Patmowt'iwn Alowanic' asxarhi (19,930 B), Girk' Tlt'oc' extracts (220,198 B), Mxit'ar Herac'i Consolatio febrium (28,369 B), Patmowt'iwn vrac' = Historia Iberorum in Armenian (5,496 B - a short extract only), Zamagirk'/Saraknoc' (14,933 B), Physiologus Armeniaca (Muradyan's edition, 15,697 B), Frik's Diwan (90,959 B), and the Corpus textuum hagiographicorum armeniacorum (40,970 B first page). OLD GEORGIAN - 185 open text units across 182 corpus directories: Kartlis Cxovreba vol. I (kcx1001-..., 113,383 B for the first chapter page, 5,522 Georgian codepoints sampled), Old Georgian Four Gospels in Khanmeti and Adishi redaction (Sarjveladze/Tvaltvadze 2001-02), Old Georgian OT from the Oshki and Jerusalem Bible manuscripts, Psalter Recensio A (Sanize 1960), Sinai Lectionary (Khanmeti, Graz ms + BnF georgien 30), Palimpsest Gospel Codex A89/A844 (Kajaia 1984), hagiography vols I-VI, Sinai hagiographical texts, Ephrem Syrus sermons (351,523 B), John Climacus Scala (Poti 1902), Satberdi collection (10th c., 140,726 B), Georgian Synaxarion (Marr, PO 19/5, 1926), liturgies (James, Chrysostom), Nomocanon, Mikael Modrekili hymnography, plus 401-gated Gregory of Nazianzus versions and the Vienna palimpsest. MANUSCRIPT METADATA: 2,580 Armenian manuscripts in 50 digital libraries (2,309 with a IIIF manifest string, of which 528 are open http manifests and 1,781 say 'Log-in required'; 2,565 carry a permalink; biggest holders HMML 1,780, Gallica 182, Vatican 138, Bodleian 73, UCLA 63, Staatsbibliothek Berlin 51, Library of Congress 47, IAE Yerevan 36, IRHT/ARCA 35, Polona 25, Morgan 13; genres Liturgy 639, Bible 385, Theology 305, Exegesis 126; 245 rows are Matenadaran shelfmarks, 234 with a manifest). ARMENIAN TEXT EDITIONS AS PDF: 115 PDF links on Matenadaran's digital-library page (Sebeos, Koriwn, Shnorhali, Hovhan Damaskaci vol. II, Kirakos Gandzaketsi, Samuel Anetsi, Girk' Tght'oc' A-C, Erznkatsi, Narekatsi studies, Latin, Hebrew and 'Parsic' catalogues) plus the Mayr C'uc'ak Grand Catalogue vols 1-11 (Hator-1.pdf 9.2 MB ... Hator-11.pdf 20.4 MB) and the concise catalogue vols 1-3 (Dzeragrats-1.pdf 104.6 MB). OPEN WIKITEXT: Armenian Wikisource holds 3,060,684 bytes across the pages I measured - Xorenac'i Patmowt'iwn Hayoc' 641,160 B, Agathangelos in 4 parts (60,080 + 215,200 + 167,640 + 446,733), Narekac'i Matean Olbergut'ean 763,209 B, Koriwn Vark' Mastoc'i (Abeghyan ed.; 31 wiki pages = index + title page + chapter pages), and the 1896 'Ararat' Armenian Bible (Genesis 304,756 B, Exodus 255,093 B, Leviticus 183,030 B, Ruth 19,744 B - the other six books linked from the index are red links). PALIMPSESTS: 74 Sinai palimpsests imaged (Arabic, CPA, Georgian, Greek, Slavonic, Syriac), of which 15 are Georgian (Georgian 10, 34, 49, NF 7, 13, 19, 55, 59, 61, 71, 84, 90, frg 68a, 72a, 73a) and none Armenian. IMAGES: 36 Armenian manuscripts (Gospels, hmayil amulets, Narek scrolls) at iae-archive.am with a 49.6 MB album PDF. NOT COVERED ANYWHERE: no open transcription/OCR of any Armenian or Georgian manuscript; OCR'd full texts exist only where TITUS or Wikisource already typed them.

Licence: not stated

Access: TITUS: no archive, no git, no OAI; Apache directory listing is 403 (/texte/etcs/cauc/ageo/ -> 403), so crawl from texte2.htm and then walk <code>NNN.htm> sequentially (armnt001..029 verified; the JS index block inside each frameset lists the chapter count). Armenian manuscripts: curl https://raw.githubusercontent.com/CVidalG/Index-of-Armenian-Manuscripts/master/index.tsv (577,642 B, 2,580 rows) or the Zenodo v3.0 file index.tsv (577,398 B, CC BY 4.0); Matenadaran: direct GETs under https://www.matenadaran.am/ftp/ (el_gradaran/*.pdf, data/M/Hator-1..11.pdf, Hamarot/Dzeragrats-1..3.pdf); Wikisource: MediaWiki API (action=query&list=allpages&apprefix=...) or Special:Export. · sample HTTP 200

Not reachable: https://cudl.lib.cam.ac.uk/search?keyword=georgian&format=json -> 403 CloudFront 'The request could not be satisfied' (Cambridge Digital Library blocks this host); canonical alternative: the IIIF/manifest route of each holding library listed in index.tsv.; https://www.loc.gov/item/00271073392-jo/manifest.json -> 403; https://www.loc.gov/item/00271073392-jo/ -> 403 Cloudflare challenge ('Just a moment...'); the JSON variant ?fo=json returned 200 once (167,922 B, item 'J Georgian 147' = a Georgian manuscript) and then 403 on repeat - LOC access from here is rate-limited/blocked, so LOC's 47 Armenian + Georgian items are only reachable in bulk through the Index TSV.; https://gallica.bnf.fr/iiif/ark:/12148/btv1b100933677/manifest.json -> 403 (Gallica blocks this host; already recorded in the gallica-bnf dossier). 181 Armenian manuscripts in the Index point at Gallica.; https://geomanuscript.ge/ and http://geomanuscript.ge/ -> TCP connect timeout (DNS resolves to 91.239.207.211, no HTTP response after 25 s); the National Centre of Manuscripts' Georgian-manuscripts portal is unreachable from this machine. Canonical alternative: manuscript.ge itself (200) for descriptions, TITUS for the transcribed Old Georgian texts.; https://sinaipalimpsests.org/ and https://www.sinaipalimpsests.org/ -> TLS handshake fails (curl exit 000); only the plain-http http://sinaipalimpsests.org/ answers 200. The research site (https://sinai.library.ucla.edu/browse) answers 200 with 'Unauthorized - Please login to access this page'.; https://www.armenianbible.org/ -> 200 but it is a parked Hostinger domain ('Parked Domain name on Hostinger DNS system'), not a Bible text. armenianbible.net and bible.armenianchurch.org do not resolve.; https://matenadaran.am/sitemap.xml -> 301 with an empty body; /sitemap_index.xml -> 404. Only /wp-sitemap.xml works. There is no IIIF service and no online manuscript viewer at matenadaran.am: 'Online Manuscripts' just links out to armenian-manuscripts-index.com, and digitisation is a paid service ('Digital Copies' page).; TITUS directory listings (/texte/etcs/cauc/ageo/, /texte/etca/cauc/, /texte/etcs/arm/) -> 403 Forbidden, so corpus membership cannot be enumerated by listing; it must come from texte2.htm.; 7 Old Georgian and 2 Old Armenian entries -> 401 (members-only): the six Gregory of Nazianzus versions (etcc/cauc/ageo/grnaz/*), the Vienna cod. georg. 2 palimpsest (etcc/cauc/ageo/xanmeti/vienna/vienn.htm), the Kunzle/Weitenberg Four Gospels (etcc/arm/armntbk) and the SSGG dialect recordings (etce/cauc/armen/armssgg).; hy.wikisource Bible index: 6 of the 10 linked books (ԹՈՒԵՐ, ԵՐԿՐՈՐԴ ՕՐԷՆՔ, ՅԵՍՈՒ, ԴԱՏԱՒՈՐՆԵՐ, and the remaining prophets/epistles) are red links - action=query reports 'missing'. Only Genesis, Exodus, Leviticus and Ruth have text.; ka.wikisource 'ქართლის ცხოვრება' is an index table whose 12 component-work links all return 404 (they are red links); Georgian Wikisource currently holds no chronicle text, so TITUS kcx1 is the only open transcription.; https://iae-archive.am/eng/scripture/38/1/ and the /eng/scripture/... permalinks stored in index.tsv -> 404; the working routes are the numbered WP paths (/eng/37/Home, /arm/41/...). There is no IIIF and no per-manuscript PDF link - only the album.

1) TITUS: (a) parse https://titus.uni-frankfurt.de/texte/texte2.htm, sections delimited by <A NAME="aarmen"> (Old Armenian) and <A NAME="georgant"> (Old Georgian); each entry is <LI><A NAME=...><B>title</B> ... <A HREF=...>; take the HREF, not the neighbouring /personal/gippertj.htm link - a naive…

biblissima-iiif: Biblissima IIIF Collections — a cross-collection search engine over harvested IIIF manifests of manuscripts and rare books dated before 1800 (Biblissima+, Campus Condorcet / Équipex Biblissima)

Counted live with q=* and per-collection filters: 144,418 manifests in 37 collections. Largest: Gallica (Bibliothèque nationale de France) 59,326; Arca (IRHT) 39,503; Bayerische Staatsbibliothek 11,916; Digital Bodleian 5,882; Teca digitale – Biblioteca Medicea Laurenziana 4,125; e-codices 2,127; Digital Collections (Leiden) 1,144; The British Library, Polonsky Pre-1200 Project 405. Other collections include Cambridge Digital Library, Parker Library on the Web, Manchester Digital Collections, Durham, Huntington, Harvard, Toronto (Fisher), Heidelberg (Bibliotheca Palatina), FulDig, Ghent, Biblioteca Nacional de Portugal, Europeana Regia, Mazarinum, NuBIS, Rosalis, Rotomagus, PaGella, Mémonum, BVH, CBMA, Cecilia (Albi), Mmmonk, ENSBA, Ludwig-Maximilians-Universität München, Patrimonio Digital Complutense, Staats- und Universitätsbibliothek Bremen, Badische Landesbibliothek Karlsruhe, Bibliothèque de l'… (truncated in the page markup), Arca. Single-token hit counts: 'Anastasius Sinaita' 222, 'Ephraem' 213, 'Photius' 94, 'Nilus' 90, 'Cassian' 239, 'Symeon Stylites' 138, 'Choniates' 25, 'Neophytos' 17, 'Amphilochia' 2. The Arca/IRHT manifests are the most useful: they carry a 'Auteur, titre, oeuvre' metadata field naming the text (e.g. 'LEONTIUS NEAPOLEOS CYPRI ep., Vita S. Joannis Eleemonis'), which is what the search indexes.

Licence: not stated

Access: No dump. 144,418 manifests are indexed; enumerate by 37 collection filters, 20 hits per search page, then read manifests directly from the providers (no rate limit observed at ~1 req/s). · sample HTTP 200

Not reachable: https://portail.biblissima.fr/ — Anubis proof-of-work challenge ('Making sure you're not a bot!'), needs JavaScript; iiif.biblissima.fr is the unblocked surface; https://gallica.bnf.fr/iiif/ark:/…/manifest.json — HTTP 403 from this host for every ark tested (btv1b10501706q, btv1b10723121w); the arks are still valid identifiers

1) Query /collections/search with a single token (author name or a distinctive word of the work title) and optionally &collection=… to scope; parse data-shelfmark and data-manifest from the result list. 2) Fetch each data-manifest URL directly with a browser UA (Gallica 403s, Laurenziana 302s — use…

ccel: Christian Classics Ethereal Library (ccel.org) - Schaff ANF/NPNF and 1,371 English works, with 7 Ancient Greek items

Counted from https://ccel.org/index/language on 2026-11-14: English 1,371 works; Ancient Greek (to 1453) 7; Latin 5; Russian 17; German 4; French 3; the rest single figures. The 7 'grc' items are 6 Greek New Testament editions (Scrivener 1894, Stephanos 1550, Westcott-Hort, SBL GNT, SBL GNT with editorial symbols, a modern-Greek Bible) plus Kirsopp Lake's Apostolic Fathers (Loeb 1912, Greek + English). The patristic substance is the Schaff series in English: Ante-Nicene Fathers 10 vols, Nicene and Post-Nicene Fathers series 1 (14 vols) and series 2 (14 vols) = 38 volumes, each also offered as PDF/ePub/TXT. That covers, for example, Basil's De Spiritu Sancto, the nine Homilies on the Hexaemeron and the Letters (NPNF2-08), and John of Damascus, Exposition of the Orthodox Faith (NPNF2-09).

Licence: not stated

Access: Per-volume downloads only. 38 ANF/NPNF volumes x 4 formats is a small, tractable set; there is no full-corpus archive and no OAI/REST endpoint. · sample HTTP 200

CCEL is an English-translation oracle, not a Greek-text source. Use it in two ways. (a) Translation-status verification: fetch /ccel/schaff/npnf1<NN> or npnf2<NN> (volumes 01-14 each) and grep the volume contents for the Latin/Greek incipit; the series' own 'cache/<work>.txt' is the cleanest full t…

CMCL — Corpus dei Manoscritti Copti Letterari, 'Letteratura Copta' edition server (Unione Accademica Nazionale / CIM Roma)

Coptic literary texts (Sahidic unless noted), Clavis Coptica-numbered, edited by Tito Orlandi and collaborators. Counted live 2026-11-14: top level 36 files = 20 HTML + 16 PDF; testi/ = 17 work directories / 679 files (PDF editions, JPG/JPEG folio images, TTF Coptic fonts, grammar and word indexes); rassegna/ = 16 PDFs. Works present in testi/: cc0223 De anima et corpore (ps.-Athanasius/Alexander/Melito/Chrysostom), cc0023 Antony Epistulae 4-6, cc0065 Vita Antonii (ed. Sheridan), cc0407 Vita Aphou, cc0127 Damianus Alexandrinus De Nativitate (86 files incl. per-folio HTML transcription), cc0108 Cyril In Athanasium, cc0415 + cc0500 John of Lycopolis texts, cc0203 Historia Horsiesi (synoptic edition), cc0013 Agathon of Tarsus De incredulitate, cc0010 Contra Iustinum Samaritanum, cc0432 Passio Mercurii + cc0001/cc0002/cc0078, cc0208 Pistis Agathonikou, cc0076 Basil De templo Salomonis, cc0637 Shenoute Contra Origenistas, cc0012 Contra Stratonicum, cc0152 Vita Pauli de Tamma, plus a multimedia Historia Ecclesiastica experiment (cc0002). Standalone PDFs include Severiano di Gabala In Apostolos (Clavis Coptica 0331 / CPG 4281, ed. D. Righi, 2 vols, ISBN 88-85354-07-6) and Orlandi, Coptic Texts relating to the Church Canons (2016, ISBN 978-88-85354-20-3).

Licence: not stated

Access: recursive wget of the autoindexes https://www.cmcl.it/~cmcl/ + /testi/ + /rassegna/ (36 top-level files, 679 files under testi/, 16 PDFs under rassegna/). Serial requests only, one mirror. · sample HTTP 200

Not reachable: http://rmcisadu.let.uniroma1.it/~cmcl/ — 403 (the legacy address printed in every CMCL page header; the live mirror is https://www.cmcl.it/~cmcl/); https://www.cmcl.it/~cmcl/testi/antlet/antlet_1.pdf — 404 (numbering starts at antlet_2.pdf; use https://www.cmcl.it/~cmcl/antlet.html to enumerate)

1) Crawl https://www.cmcl.it/~cmcl/ and /testi/ by autoindex parse; pages are ISO-8859-1 — set the encoding before reading (Python requests guesses wrong). 2) Two distinct text carriers: (a) HTML editions such as testi/sheori/cc0637.html embed Orlandi's 'delta' encoding — page anchors '(p 100.00)',…

coptic-mss: Coptic manuscript/papyrus repositories, second sweep - Digital Bodleian Coptic collection (MS. Copt. papyri, MS. Bruce 96, MS. Huntington 365), Berliner Papyrusdatenbank / Aegyptisches Museum und Papyrussammlung Berlin (Coptic papyri and ostraca), Duke Papyrus Archive (Coptic literary and documentary papyri), Claremont Nag Hammadi Archive (photographs of the Nag Hammadi codices)

Counts below are enumerated, not estimated, all on 2026-11-14 UTC. (1) DIGITAL BODLEIAN 'Coptic Manuscripts and Papyri' collection: 54 items / 584 canvases. 52 are papyri with the shelfmark pattern MS. Copt. <a-g>. <n> (P) (158 canvases total; most are single-sheet recto+verso, the largest is MS. Copt. c. 19 (P)/1-23 at 46 canvases, then c. 16 (P)/1-4 at 8 and d. 12 (P)/1-2 and e. 16 (P)/1-2 at 4 each). Two are codices: MS. Bruce 96 (158 canvases, f. 148, titled 'Two Gnostic Treatises; First and second books of Ieou' - i.e. the Bruce Codex / Papyrus Brucianus, the only copy of the Books of Jeu) and MS. Huntington 365 (268 canvases, f. 128, al-Samannudi's Muqaddimat al-sullam and al-Sullam al-kana'isi, Bohairic grammar + dictionary with Arabic, 13th-c. author). The 52 (P) items are NOT described in their manifests: only the two codices carry a Title (52 of 54 do carry a 'Catalogue Description' link to the ACEA catalogue). (2) BERLINER PAPYRUSDATENBANK: 630 Coptic-bearing records out of ~9,486 described objects (facets: dokumentarisch 8,091, halbliterarisch 227, literarisch 1,168, literarisch (christlich) 245). The Coptic slice is dominated by Thebes/Diospolis ostraca and is mostly documentary; 151 records are 'Koptisch Sahidisch' + literary, of which 30 match 'Psalm'. Every record carries the shelfmark (P. <n>), material, form, line count, script description, date, provenance, a content summary with a Coptic incipit, and the publication reference - e.g. P. 740 cites Delattre, 'Ostraca des Musees de Berlin portant des extraits des Psaumes en copte', in: Lepper (ed.), Forschung in der Papyrussammlung, Berlin 2012, 387-398, no. 2. (3) DUKE PAPYRUS ARCHIVE Coptic section: 298 record links (58 literary, 5 subliterary, 235 documentary across 12 genre/century pages), against the editorial statement on /papyrus/texts/language.html that Duke 'holds a large number of sometimes quite important Coptic texts ... fourth to the tenth century'. Duke's Coptic literary list is unusually rich for biblical fragments (1 Kings, Isaiah, Matthew, Mark incl. the 'shorter ending', Luke, Acts) plus martyrdoms, homilies and liturgical leaves, most with a published edition named in the record. (4) CLAREMONT NAG HAMMADI ARCHIVE: 3,474 items, 2,374 of them papyrus/codex photographs, including leaf-by-leaf negatives of the Nag Hammadi codices taken by Jean Doresse and Basile Psiroukis at the Coptic Museum Cairo and published in the Brill facsimile edition (item 1234 cites 'The facsimile edition of the Nag Hammadi Codices [9/10] Codices IX and X, Leiden: Brill, 1977, p. 11/13'). This is the practical substitute for the Coptic Museum Cairo, whose own site is unreachable.

Licence: not stated

Access: None of the four offers a dump, and all four are cheap to enumerate without one. Bodleian: 1 collection request + 1 manifest per item (54) + 1 image request per canvas (584 canvases total; 158 for the 52 papyri, 158 for Bruce 96, 268 for Huntington 365). BerlPap: paginate /result/?Sprache=Koptisch (40 pages x 10 for the main bucket, 17 pages for Sahidic) then fetch record pages; images are one GET per side (recto/verso), the Original JPEG at ~1.6 MB is 600 dpi and needs no post-processing. Duke: 15 index pages, 298 record pages, images are 150-dpi GIFs (615x1707 for P.Duk.inv. 797 recto, 712,700 B) - no bulk zip. Claremont: CONTENTdm paging at maxRecords/100 (35 pages for the whole nha set), images are JP2 masters served as JPEG derivatives. Etiquette: Bodleian and BerlPap answered ~5-10 req/s without throttling, but Bodleian's terms ask for a credit line and BerlPap's Nutzungshinweise restrict publication, so keep a 1 req/s crawl and cache. · sample HTTP 200

Not reachable: https://quod.lib.umich.edu/a/apis and /a/apis?q1=coptic and /cgi/i/image/image-idx?q1=coptic - HTTP 403 with a 5.6 KB block page (University of Michigan Library denies this host, not a robots restriction). https://lsa.umich.edu/papyrology also 403 and https://deepblue.lib.umich.edu/ 403; https://digitalcollections.library.umich.edu and https://iiif.lib.umich.edu do not resolve (curl 6). Michigan's Coptic papyri (APIS/quod) could not be opened at all. Canonical alternative: the same APIS records are served by Papyri.info (itself Anubis-walled here) and the Kelsey/APIS metadata is mirrored in Trismegistos (403 here); otherwise use the printed editions the APIS records cite.; https://papyri.info/ - HTTP 200 but the body is an Anubis proof-of-work challenge (Making sure you're not a bot!, anubis_version v1.27.0, difficulty 5); no record content without executing JS. Canonical alternative: papyri.info identifers are printed in the BerlPap Kataloge field, and DDbDP text is embedded in BerlPap for Greek items.; https://acea.bodleian.ox.ac.uk/ and /catalog/manuscript_378 - HTTP 200 but Anubis v1.28.1 challenge page. This is the full catalogue of the Bodleian's Coptic papyri, so the 52 (P) items cannot be identified automatically. Canonical alternative: the IIIF manifests (titles for the two codices only) and the printed Bodleian Coptic papyri catalogue.; https://digital.bodleian.ox.ac.uk/search/?q=coptic - HTTP 200 but a 10,980-byte SPA shell with zero results; ?format=json returns the same shell; /api/search/ is 404. No keyword search is scriptable. Alternative: enumerate https://iiif.bodleian.ox.ac.uk/iiif/collection/coptic (54 items).; https://onb.digital/ and https://onb.digital/search?q=koptisch - HTTP 504 Gateway Timeout (24-byte text/plain body) on every attempt; https://digital.onb.ac.at/OnbViewer/viewer.faces - connection timeout (curl 28, 30 s). The Austrian National Library's discovery layer for the Papyrussammlung is therefore unusable from this host, although the API itself is up: https://api.onb.ac.at/ (200) documents IIIF Presentation 3.0 (/iiif/presentation/v3), IIIF Image 3.0 (/iiif/image/v3), IIIF Content Search 2.0 (/iiif/search/v2) and REST endpoints GET /rest/v1/ie/<ieId>/{file|alto|txt|preview|thumbnail}, i.e. per-page ALTO and plain text. https://viewer.onb.ac.at/ (200) is a React SPA whose MANIFEST_URL is https://api.onb.ac.at/iiif/presentation/v3/manifest/<ieId>. Without a reachable catalogue I could not obtain a single ieId, so no Vienna Coptic item is listed here. Canonical alternative: the ÖNB Papyrussammlung states at https://www.onb.ac.at/sammlungen/papyrussammlung/bestaende that 'Die umfangreichsten Bestaende liegen in arabischer, griechischer und koptischer Sprache vor', and Vienna's Coptic papyri are catalogued in the printed P. Vindob. series and indexed in Trismegistos (403 here).; https://digitalisate.sub.uni-hamburg.de/ - 302 then a BunkerWeb captcha wall (Bot Detection / BUNKERWEB / SECURITY CHECK / CAPTCHA / MANUAL CODE REQUIRED). Hamburg's digitised manuscripts cannot be searched or fetched. https://www.qalamos.net/ (and the bare host, and /api/v1/search) - connection timeout (curl 28, 15-25 s), so the Qalamos portal that now carries the KOHD records for German collections is unreachable; https://kohd.uni-goettingen.de does not resolve (curl 6). https://handschriftenportal.de/ answers 200 with a 951-byte JS shell and no API (/api is the same shell); https://api.handschriftenportal.de does not resolve. Hamburg/Marburg/Berlin Coptic manuscript records remain a blind spot; canonical alternative: Qalamos from a network that can reach it, or the printed KOHD volumes (Verzeichnis der Orientalischen Handschriften in Deutschland), which BerlPap cites as 'KOHD Coptica (Hasznos)'.; https://orient-digital.staatsbibliothek-berlin.de/ - connection timeout (curl 28, 25 s). The Staatsbibliothek zu Berlin's Oriental digitisation portal is not reachable, although https://content.staatsbibliothek-berlin.de/ (200, 221 B), https://digital.staatsbibliothek-berlin.de/ (200, 4,125 B SPA shell) and https://staatsbibliothek-berlin.de/ (200) answer. Berlin's Coptic manuscripts (as opposed to the papyri in BerlPap) could not be enumerated.; https://collections.louvre.fr/en/recherche?q=copte - HTTP 429 with a 4.6 KB bot wall (WADP - I am not a robot); the Louvre's IIIF prefix still cannot be determined. The Louvre holds Coptic textiles, stonework and the Bawit frescoes rather than Coptic literary manuscripts (the manuscripts are at BnF), so this is a low-value gap for texts. Canonical alternative: none for images; the objects are published in the Louvre's printed catalogues.; https://www.copticmuseum.gov.eg/ - connection timeout (curl 28, 18 s). The Coptic Museum, Cairo (owner of the Nag Hammadi codices) has no reachable online catalogue. Practical substitute found and verified: the Claremont Nag Hammadi Archive photographs of the same Cairo leaves (see new_texts).; https://gallica.bnf.fr/* - still 403 (Cloudflare WAF), consistent with the existing gallica-bnf dossier; https://archivesetmanuscrits.bnf.fr/ likewise unusable for record lists. BnF Copte manuscripts (Copte 1-103+) remain image-blocked here.; https://www.papyrology.ox.ac.uk/POxy/ - TLS failure: no alternative certificate subject name matches target host name (curl 60), on both the home page and the search page. Oxyrhynchus Online could not be checked for its Coptic items.; https://paths.adwmainz.net/ - TLS handshake failure (tlsv1 alert internal error, curl 35), unchanged from the first sweep; PAThs (the natural work-to-manuscript authority for Coptic literature) is still unreachable.; https://repository.duke.edu/dc/papyri - HTTP 200 but Anubis challenge page; only the legacy static archive (library.duke.edu/papyrus/) is usable, which means no higher-resolution Duke images.; https://collezioni.museoegizio.it/ - HTTP 200 (148 KB) but the search is client-side; ?q=copto returns a shell whose embedded state shows Risultati: [] and no result markup, and no public JSON endpoint was found. Turin's Coptic holdings (mostly textiles) were not enumerable in the time available.

FOUR PIPELINES, all verified end to end today. (A) DIGITAL BODLEIAN (best rights, best images). 1) GET https://iiif.bodleian.ox.ac.uk/iiif/collection/coptic once; read .manifests[] - each member gives @id (manifest URL) and label (shelfmark). Do not try to crawl the website search: it is a Vue SPA…

documenta-catholica-omnia: Documenta Catholica Omnia (Cooperatorum Veritatis Societas) - Migne PG/PL and allied collections as per-work PDFs

Counted by parsing the 20 chronological index pages on 2026-11-14: 16,252 distinct per-work file links. By filename language tag: MLT (Migne Latin) 6,599; LT 2,630; EN 1,269; IT 1,117; MGR (Migne Greek, per work) 690; GM (Greek whole-volume) 425; FR 306; ES 214; GR (non-Migne Greek edition) 156; PT 96; RU 26; RO 6; PL 5; AR 2; 2,711 untagged. Greek-language total: 1,271 files. 425 files carry the GM (Greek Migne) tag: 167 of them are whole-volume PDFs named 'PG_<vol>_(nn-00)_<Author>_Opera_Omnia__GM.pdf' and cover 160 of the 161 numbered volumes (no whole-volume file for PG 30); the rest are column-range slices such as '...(MPG_102_0015_0264)__GM.pdf', which is why the per-work MGR count (690) understates what is actually present. Beyond Migne the same index lists Patrologia Orientalis (Graffin-Nau), Corpus Scriptorum Historiae Byzantinae, Acta Sanctorum, Mansi, Miklosich-Mueller Acta Graeca, Hefele, Fabricius and Muratori. English files (1,269) are text PDFs converted from CCEL's Schaff/NPNF pages - the PDF for John Chrysostom 'De diabolo tentatore [Schaff] EN' prints 'this file has been downloaded from http://www.ccel.org/ccel/schaff/npnf109.html' in its first lines.

Licence: not stated

Access: Technically yes, legally no. HTTP Range works on the PDFs (206, application/pdf, Accept-Ranges: bytes). But the site's Terms forbid harvesting: 'The use of download sustained by spreaders and/or robots is strictly forbidden' and 'No more than 10% of any given low-level section can be downloaded and stored on local hard disk'. No mirroring, no modification of texts/properties, no password forcing. · sample HTTP 206

Not reachable: https://www.documentacatholicaomnia.eu/ -> TLS error 60 (server omits the Let's Encrypt intermediate; chain of length 1). The site is fully usable over http://, which also returns 200 for the sample PDF.

Rights first: the ToS forbids robots and caps downloading at 10% of a section, so this is a *verification and scan-fallback* source, not a harvest target - do not bulk-fetch it, and do not republish its files. Practical route for a single work: (1) locate the wrapper page by grepping the 20 a_1011_…

europeana: Europeana — the EU cultural-heritage aggregator (API v2). Not a text source in itself: it exposes provider metadata, per-item rights statements and the provider's viewer/IIIF for manuscripts from libraries the project does not otherwise index (Chester Beatty, National Library of Israel, Cyprus, Poland, Romania)

63,132,490 items across all media (totalResults for query=*). Query counts measured today: 'Theodoret' 349; 'Anastasius Sinaita' 77; 'Photius' 324; 'Ephraem Graecus' 111; 'coptic' AND manuscript 615; 'syriac' AND manuscript 85; 'geez' 68; 'ethiopic' AND manuscript 3. Provider-scoped: DATA_PROVIDER:"Chester Beatty" = 4,013 items, of which 473 match 'coptic' — that is Coptic literary and documentary material (Sahidic and Subachmimic) with IIIF at viewer.cbl.ie, a collection absent from every source in the current register. Language filtering is weak: qf=LANGUAGE:cop / LANGUAGE:gez / LANGUAGE:"Sahidic (Coptic dialect)" all return 0, so language must be inferred from dcLanguage per item.

Licence: not stated

Access: No bulk download without a key. Paging by rows/start works; 63,132,490 items overall, so always scope with qf (DATA_PROVIDER / LANGUAGE / TYPE) before paging. · sample HTTP 200

Not reachable: https://www.europeana.eu/en and https://pro.europeana.eu/page/terms-of-use — HTTP 403 Cloudflare interstitial from this host; the API (api.europeana.eu) answers normally; https://www.europeana.eu/item/1100/6401 — item pages sit behind the same 403; use the API record endpoint or the provider URL in edmIsShownAt

1) Always scope with qf before paging: qf=DATA_PROVIDER:"Chester Beatty" cuts 63M items to 4,013. 2) Read edmIsShownBy: for CBL it is a JPEG under viewer.cbl.ie/viewer/content/<id>/800/0/<file>.jpg, and the IIIF manifest is https://viewer.cbl.ie/viewer/api/v1/records/<id>/manifest/ — note the trail…

Die griechischen christlichen Schriftsteller 13 (GCS) — Carl Schmidt, Koptisch-gnostische Schriften I (Leipzig, Hinrichs 1905), Internet Archive copy

One volume, three works, Coptic text with German translation, from the two great pre-Nag-Hammadi Gnostic codices: (1) Die Pistis Sophia (Codex Askewianus, BL Add. 5114); (2) Die beiden Bücher des Jeû (Codex Brucianus, Bodleian MS. Bruce 96); (3) Unbekanntes altgnostisches Werk (the 'Untitled Text' of the Bruce codex). Table of contents verified in the text: 'DIE PISTIS SOPHIA — DIE BEIDEN BÜCHER DES JEÛ — UNBEKANNTES ALTGNOSTISCHES WERK' with part divisions at pp. 1, 257, 335. Full-text-inside counts: 'Pistis Sophia' 364 matches, 'Jeû' 124, 'Unbekanntes' 26. Overlap note for the merge: the sibling file sources-more-ia-gcs.json covers the same series' GREEK volumes (GCS 8, 21, 25, 31, Origenes Werke, Eusebius Werke, Clemens, Hippolyt). GCS 13 (this dossier) is the Coptic volume — same publisher and series, different language and different corpus; count both, but cite the series once.

Licence: not stated

Access: PDF https://archive.org/download/koptischgnostis00schm/koptischgnostis00schm.pdf (39,594,242 bytes); jp2 zip; *_djvu.txt (Range requests work). · sample HTTP 200

Not reachable: https://www.bl.uk/manuscripts/Viewer.aspx?ref=add_ms_5114_f001r — 200 but serves the generic 'Digitised manuscripts and archives' landing page, which states 'some digitised items are currently unavailable as a result of a cyber-attack on the Library'. The Askew Codex images are not reachable; canonical alternative for the text is this GCS volume.; https://digital.bodleian.ox.ac.uk/search/?q=bruce+96 — the search is JavaScript-only (10 KB shell, no results) and https://digital.bodleian.ox.ac.uk/collections/coptic/ returns 404; no IIIF manifest for MS. Bruce 96 was located. Alternative for the Bruce codex text: this GCS volume.; https://digital.bodleian.ox.ac.uk/objects/ — 404; the object path pattern could not be discovered without executing JavaScript.

1) Two PD copies; use koptischgnostis00schm (IA/Toronto) because it has a full jp2 zip and 23 files; the Google copy has thinner derivatives. 2) Take page images with /download/<id>/page/n<N>.jpg and map N to the printed page using /download/<id>/<id>_page_numbers.json if present, or the full-text-…

The Gnostic Society Library — Nag Hammadi Library (gnosis.org), English translations of the thirteen codices

All thirteen Nag Hammadi codices in English. Site text: 'a collection of thirteen ancient books (called "codices") containing over fifty texts'. Counted 2026-11-14 from nhlalpha.html: 99 translation links resolving to 41 distinct titles on the A-Z page (several works carry 2-4 rival translations, e.g. Gospel of Thomas has Lambdin CGLP, Meyer, Davies, Patterson & Meyer, Barnstone; Apocryphon of John has Wisse CGLP, Waldstein & Wisse short and long, Meyer, Davies). Codex index nhlcodex.html organises the same texts by NHC I-XIII. No Coptic text anywhere on the host.

Licence: not stated

Access: wget -r -np http://gnosis.org/naghamm/ (all *.html; ~99 translation links plus introductory essays). · sample HTTP 200

Not reachable: https://gnosis.org/naghamm/nhl.html — TLS failure (self-signed + expired certificate, verify code 10). The canonical, working URL is http://gnosis.org/naghamm/nhl.html (200).; http://gnosis.org/library/pistis-sophia/index.html — 404; the index lives at http://gnosis.org/library/pistis-sophia/ (200).; https://brill.com/display/db/COGN — 405 CAPTCHA; Coptic Gnostic Library Online cannot be read from this machine.

1) Use http:// not https:// — the site's TLS certificate is self-signed and expired (openssl: 'verify error:num=18:self-signed certificate', 'num=10:certificate has expired', subject/issuer CN=gnosis.org), so requests/curl fail with CERTIFICATE_VERIFY_FAILED; plain HTTP returns 200. 2) Fetch nhlalp…

goarch: Greek Orthodox Archdiocese of America (goarch.org) and its Digital Chant Stand (dcs.goarch.org)

Digital Chant Stand only: 8,523 unique score/service PDFs in 28 collections, ~2,145 tagged 'gr'. The 'b'/'w' path segment distinguishes book/webscreen variants. These are Byzantine chant scores (neumes plus text) and service booklets for the daily cycle, not an e-text library of patristic works: the sample checked is a 7-page score of the Annunciation canon. The Archdiocese's actual text pages (e.g. /chapel, the online Chapel texts, the encyclicals) are unreachable from here, so no count can be given for them.

Licence: not stated

Access: booksindex.html contains 8,523 unique PDF URLs of the form https://dcs.goarch.org/media/m/<collection>/<lang1>/<lang2>/<mode>/<b|w>/<service>.pdf (10,832 hrefs before de-duplication). Language pairs: en/me 2,996, en/oc 966, en/pe 952, en/tr 895, gr/oc 710, gr/tr 659, gr/me 437, gr/pe 339 -> about 2,145 Greek ('gr') score PDFs. Collections: dedes 5,835, velon 653, theodoridis 491, apostolou 367, newbyz 257, jmboyer 186, roubanis 147, cremeens 128 (28 collections in all). · sample HTTP 200

Not reachable: https://www.goarch.org/ -> 403 (Cloudflare 'Just a moment...' challenge); same for /chapel. No page of the Archdiocese's own site could be fetched from this host.; https://www.goarch.org/chapel -> 403, same cause.

Do not plan ingestion from goarch.org: the main site 403s behind Cloudflare and the only reachable host serves scores, not texts. If the liturgical text itself is needed, better routes are: (a) the Digital Chant Stand's own HTML service pages once you resolve the JS calendar (require.js modules und…

NTVMR — Virtual Manuscript Room (VMR) of the Institut für neutestamentliche Textforschung (INTF), University of Münster: Greek New Testament manuscript catalogue, image index and TEI transcription service

Greek New Testament manuscripts — the INTF's Kurzgefasste Liste, digitised and partly transcribed. MEASURED 2026-11-14 against the live API. Catalogue: 7,771 manuscript records (metadata/liste/get, 'manuscripts.count' = 7771; unique docIDs 7,771, distinct Gregory-Aland numbers 7,266, i.e. 505 numbers are split across dispersed shelf records); 557 holding institutions (metadata/institute/getnames); 6,499 records carry a shelf instance; 2,112,502 pages catalogued in total. Images: 5,219 records have images (1,984,038 pages). Transcriptions: 908 records have at least one PUBLISHED transcription (240,090 indexed pages) — 11.7% of the catalogue. Breakdown of the nine repositories in the brief, all present (docCount / pages / transcribed manuscripts): Sinai, Saint Catherine's Monastery (instID 6) 346 / 111,611 / 29; Athos Great Lavra (113) 260 / 120,244 / 28; Athos Vatopedi (120) 197 / 84,865 / 23; Athos Iviron (108) 106 / 44,048 / 8; Patmos, St John the Theologian (175) 85 / 38,469 / 6; Jerusalem Patriarchate (263) 170 / 64,833 / 7; Biblioteca Apostolica Vaticana (460) 403 / 131,542 / 53; BnF Paris (67) 420 / 156,421 / 65; Bayerische Staatsbibliothek München (47) 33 / 15,185 / 5; Biblioteca Medicea Laurenziana (272) 79 / 33,569 / 19; Biblioteca Ambrosiana (281) 58 / 19,362 / 7. Whole of Mount Athos: 25 institutions, 1,071 records, 442,679 pages — the largest single Athos manuscript index reachable from this machine, and it is only a by-product of the NT register. Greece overall: 2,008 records, 138 with transcriptions. Also present and useful beyond the NT: lectionary records (L-numbered, e.g. L1178 = docID 41178, Patmos), and the dispersed-manuscript records that tie one GA number to several institutions at once (GA 102 = BnF Suppl. gr. 1155 ff.1-2 + Protaton 56 + Vatopedi 1219; GA 64 = Sinai Harris 10 + Sinai Syr. 7 + NLR St Petersburg Gr. 276 + Vernadsky NL Kyiv). WHAT IS ACTUALLY NEW FOR naxiai.com/christian: First1KGreek holds the Greek NT only as Mark 1:1-45 (827 words) and Perseus holds one eclectic text (Westcott-Hort) per book with the World English Bible beside it; this source supplies 908 INDIVIDUAL WITNESSES, verse-keyed, each with its own spelling, lacunae and nomina sacra, plus 1,984,038 page images linking those witnesses to the nine repositories. It adds no new literary work — it adds the manuscript dimension the project currently has only for Coptic and Ethiopic. READ THIS BEFORE USING translation_status: every entry below is marked 'full' at WORK level, because all 27 New Testament books have a complete English translation (World English Bible, PerseusDL canonical-greekLit tlg0031.*.perseus-eng2). No translation of any INDIVIDUAL MANUSCRIPT exists in this source or anywhere else, and the TEI carries no alignment, so a manuscript-level English rendering would have to be produced by the project.

Licence: not stated

Access: No archive, no git repo, no OAI-PMH. The sanctioned bulk path is the two documented exports: (1) one GET of https://ntvmr.uni-muenster.de/community/vmr/api/metadata/liste/get/?format=csv (1,081,966 bytes as JSON / ~1 MB as CSV, header 'manuscripts.count' = 7771, then one row per manuscript with gaNum, docID, orig, origEarly, origLate, lang) gives the entire catalogue; (2) one GET per manuscript of transcript/get/?docID=<n>&format=tei&pageID=ALL. Measured: 14 manuscripts = 15,218,939 bytes of TEI (8 KB-3.9 MB each, avg 1.09 MB), 1-2 s per request. Full corpus estimate: 908 transcribed manuscripts on the same average ~1.1 GB of TEI, ~2,300 requests. robots.txt sets Crawl-delay 10 and Disallow /community/vmr/api/ for crawlers, with the note 'Bulk data consumers: please contact the INTF, or use the documented exports rather than walking the API' — so pull the two exports at ~1 request/2 s and mail the INTF for the full 908-manuscript set rather than walking docIDs. · sample HTTP 200

Not reachable: https://ntvmr.uni-muenster.de/ , /vmr-manuscripts , /manuscript-workspace , /community/modules/papyri/?site=VMR&image=20001/40/10 — HTTP 403 (199-byte Apache 'Forbidden') for these paths from this machine, with curl, with a browser User-Agent, and via the FetchURL tool (/vmr-manuscripts returned 403 through FetchURL as well). Not site-wide: /liste (200, 23,057 bytes), /robots.txt (200) and /sitemap.xml (200, 77,275 bytes) are served normally, so the block targets the JS application routes; the web UI needs a real browser session and /community/vmr/api/* is the whole machine-readable surface.; https://ntvmr.uni-muenster.de/oai?verb=Identify and /community/vmr/api/oai?verb=Identify — HTTP 404 (404 with a 22,323-byte HTML body at the first, 769 bytes at the second). No OAI-PMH, so no incremental harvest; diff the Kurzgefasste Liste instead.; https://ntvmr.uni-muenster.de/community/vmr/api/transcript/get/?docID=<n>&format=csv and ...&format=json — HTTP 200 but the body is the 4,376-byte HTML parameter-documentation page, not a transcript. Silent failure, not a block; use format=tei|xhtml.; https://ntvmr.uni-muenster.de/community/vmr/api/statistics/document/dispersal/ — hangs: no response in 60 s (killed). The per-document statistics export is unavailable in practice; metadata/liste/get replaces it.; Biblioteca Ambrosiana (https://ambrosiana.comperio.it/) — NOT blocked by NTVMR but still unreachable directly, re-tested 2026-11-14: every path returns HTTP 200 with a 16,879-byte JS challenge ('Verifica in corso / Checking'), a cookie jar set on the first request changes nothing, and no IIIF endpoint can be confirmed. The register's conclusion stands; NTVMR is the workaround for its Greek NT text.; Gallica (https://gallica.bnf.fr) — still HTTP 403 from this host; BnF manuscript images are reachable only where NTVMR proxies them (GA 04 attribution 'Images by Gallica') and via the embedded digi.vatlib.it / gallica.bnf.fr IIIF manifests, which are themselves 403 here.; No bulk archive exists: no git repo, no Zenodo deposit, no tarball. 908 transcripts must be fetched one request at a time, and robots.txt asks bulk consumers to contact the INTF first.

ACCESS MODEL. Only /community/vmr/api/* is reachable from a script; the web application returns a bare Apache 403 (199 bytes, 'You don't have permission to access this resource') for /, /vmr-manuscripts, /manuscript-workspace and /community/modules/papyri/?site=VMR&image=... — use a browser for the…

handschriftenportal: Handschriftenportal (HSP) — the German national union catalogue of book manuscripts in German collections, successor to Manuscripta Mediaevalia

158,250 manuscript records (metadata.numFound for q=*). Language facet, counted from the live facet block: lateinisch 42,702; deutsch 28,164; französisch 2,161; griechisch 1,160; italienisch 880; niederländisch 641; hebräisch 403; englisch 204; spanisch 157; kirchenslawisch 111; tschechisch 90; russisch 70; polnisch 67; arabisch 63; sorbisch 53; schwedisch 45; türkisch 36; dänisch 26; aramäisch 15; slawisch 13; chinesisch 12; isländisch 11; portugiesisch 9; persisch 6; katalanisch 5; kroatisch/serbisch/jiddisch 4 each; slowenisch/ungarisch/friesisch 3; syrisch 2; äthiopisch 2; georgisch 2; tibetisch 2; armenisch 1; 89,931 records have no language value. Digitisation: 37,926 records with digitized-object=true, 29,768 with a IIIF manifest; of the 1,160 Greek records, 666 have IIIF and 694 some digitisation. Record types: hsp:object 81,470, hsp:description_retro 42,335, hsp:description 32,538; 84,082 have a description (extern 76,056). Repositories: 257 values (München 50,269; Berlin 22,517; Wolfenbüttel 10,265; Dresden 9,546; Leipzig 5,431; ... Vatikanstadt 2,022). Content is work-level in title-display ('Drei Homilien von Michael Akominatos, griech.', 'Catena in Iob', 'Theodorus Studita: Catechesis Parva et Magna'), so the catalogue itself names the text, not just the shelfmark.

Licence: not stated

Access: No dump, OAI-PMH or SPARQL endpoint found. Paging via rows+start works (tested rows=1..30); 158,250 records total. Each record is one HTTP GET, and one extra GET per PURL if the human-readable page is wanted. · sample HTTP 200

Not reachable: https://www.manuscripta-mediaevalia.de/ — HTTP 301 to https://handschriftenportal.de/ ; the old MM interface is retired and HSP is the canonical successor; https://iiif.ub.uni-leipzig.de/0000012990/manifest.json — curl exit 000 (no connection) on repeated attempts from this host, while the HSP record for Leipzig Cod. gr. 15 (Theodorus Studita, Catechesis Parva et Magna) points at it; use the Leipzig digitisation page via the HSP PURL https://resolver.staatsbibliothek-berlin.de/HSP0005226F00000000 until it answers

1) Enumerate: GET /api/hspobjects/search with q=* and fq={"language-facet":["NORM-124c5435-5f39-3f0d-9b0f-653d105340b3"]} (Greek) or the other NORM ids resolved from /api/authority-files/{id}; page with start+=rows. 2) For each hit take hspObject.id, idno-display, title-display, repository-display,…

HTR-United — catalogue of handwritten-text-recognition ground truth (Greek and Syriac Christian manuscripts; Zenodo-hosted PAGE-XML/ALTO deposits)

Catalogued HTR/OCR ground truth for patrimonial documents: 138 datasets, 2,099,146 lines and 75,128,259 characters declared in the volume[] fields (lower bounds - some records count pages, files or regions instead of lines). Formats: ALTO-XML 95, PAGE-XML 38, image+text pairs 5. Scripts: Latn 119 datasets, Grek 12, Arab 5, Latf 4, Deva 2, Syrj 1, Syre 1, Hebr 1, Hant 1, Mlym 1. Languages of interest: grc in 10 datasets, syr in 2, ara in 4; zero records for cop, gez/amh, hye or kat - there is no Coptic, Ethiopic, Armenian or Georgian HTR ground truth in this registry (the Coptic and Ge'ez routes remain print-OCR: CopticScriptorium/OCR and HHD-Ethiopic, see adjacent_deposits). The Greek/Syriac content measured here: ÖNB Cod. Syr. 1 (140 transcribed pages, 2,871 lines, 126,407 chars, 73.8% Syriac); Jerusalem SMMJ 00036 (133 pages, 17,836 lines, 329,095 chars, 74.5% Syriac); EPARCHOS BL Add. 6791 (120 pages, 2,272 lines, 18,982 words, 112,256 chars, 80.1% Greek); Stavronikita Athos ХФ53 (54 pages, 1,038 lines / 5,592 words), ХФ114 (44 pages, 1,051 lines / 5,467 words), ХФ79 (40 pages, 803 lines / 4,389 words). The Greek manuscripts are dated 9th-16th c. and the Syriac 12th-16th c., i.e. this is the only registered source that supplies machine-readable transcription of *manuscript* (not print) Greek and Syriac Christian books that is not already in First1KGreek/Perseus (editions of texts, not MSS) or DigiVatLib (images only).

Licence: not stated

Access: git clone https://github.com/HTR-United/htr-united (repo 45.5 MB, catalogue + schema only; no manuscript text). Text/files: https://zenodo.org/api/records/{id}/files/{name}/content - 1.1-2.6 MB for PAGE-XML/ALTO zips, 230-970 MB for images.zip; Range requests are honoured (HTTP 206) so a single member can be pulled without the full archive. · sample HTTP 200

Not reachable: https://gitlab.huma-num.fr/ecrinum/anthologia/htr_cpgr23 (Ground truth for the Palatine Anthology, HTR_CPgr23, CC-BY 4.0, 50 pages / 3,374 lines / 114,273 chars per catalog.json) - TCP connect timeout after 25 s from this machine, twice; the record is real but the text cannot be fetched here. Canonical alternative: ask the ECRINUM team, or use the manuscript images (Cod. Pal. gr. 23) from Heidelberg's digitisation until the host is reachable.; https://escriptorium.huma-num.fr/ - DNS 'Could not resolve host' from this machine; https://escriptorium.fr/ fails TLS ('unrecognized name'). eScriptorium itself is not blocked as software (the deposits above were exported from it), only these public instances are unreachable here.; https://www.e-rara.ch/cgj/content/zoom/{2842014,2841606} - HTTP 200 but the body is a 'Verifying your browser' challenge, not the 1644 Dionysius print; no document retrieved.; https://spotlight.vatlib.it/greek-paleography - TLS failure (curl 60: 'unable to get local issuer certificate'), i.e. an incomplete certificate chain; a browser accepts it. This is the transcription source named in the PTA GreekHTR README, so it matters - use a browser session or a CA bundle that includes the missing intermediate.; https://transkribus.eu/Transkribus/rest/getCurrentUser, https://transkribus.eu/TranskribusRest/collections, https://api.transkribus.eu/v1/collections - all HTTP 400 without a session; Transkribus collections/models cannot be listed or exported anonymously. Open alternative: the Transkribus-produced ground truth that authors have deposited in HTR-United (the ÖNB and SMMJ 036 sets above were exported from Transkribus to PAGE-XML).; https://zenodo.org/api/records/{id} with user-agent 'Mozilla/5.0' (bare) - HTTP 403 'restricted due to unusual traffic from your network'; same URL with a full Chrome UA is 200. Pitfall, not a block.; https://zenodo.org/api/records/22098097 files - record access_right=restricted, no downloadable files despite a CC-BY 4.0 licence claim in catalog.json; access must be requested from the depositor.

1) Discover: fetch https://raw.githubusercontent.com/HTR-United/htr-united/master/catalog.json (524 KB) and select records where script[].iso is Grek/Syrj/Syre or language includes grc/syr. Each record's 'url' is either a DOI (10.5281/zenodo.NNNN) or a GitLab/GitHub URL; resolve DOIs to Zenodo reco…

ia-gcs: Die griechischen christlichen Schriftsteller der ersten drei Jahrhunderte (GCS, Hinrichs/De Gruyter critical editions, 1897-1922) on the Internet Archive — Greek text layers produced with Tesseract grc, i.e. a Migne alternative with a critical apparatus

The GCS series (Berlin/Leipzig: Hinrichs, later De Gruyter) is the standard critical edition for Greek Christian literature of the first three centuries, and it is a different corpus from Migne: new editions with apparatus, not reprints of the Patrologia Graeca. Counted by query on 2026-11-14: series-title search 39 items, identifier:diegriechischenc* 11, identifier:origeneswerke* 31, identifier:clemensalexandri* 12, identifier:eusebiuswerke* 9, title:("Hippolytus Werke") 3. After de-duplication (the same volume often has 2-4 scans) I verified 25 distinct volumes by downloading and reading their text layers: 16 carry a Greek text layer that measures as Greek (see ocr), 6 are Latin- or German-only (Eusebius Werke 2,3; 5 = German translation of the Armenian Chronicle; Origenes 6, 7, 8 = Rufinus' Latin; Eusebius 7,1 = Jerome's Latin Chronicle), 3 are apparatus/registers (Eusebius Werke 2,3; Clemens IV; Clemens 04/39). One further volume exists only as a lending copy (Eusebius Werke 8,1, 401). Authors covered with usable Greek: Origen (Contra Celsum, De oratione, De principiis, Johanneskommentar, Jeremiah homilies), Eusebius (Vita Constantini, Laus Constantini, Contra Marcellum, De ecclesiastica theologia, Demonstratio Evangelica), Epiphanius (Ancoratus, Panarion 1-64), Clement of Alexandria (Protrepticus, Paedagogus, Stromata I-VIII, Excerpta ex Theodoto, Eclogae, Quis dives), Hippolytus (Refutatio omnium haeresium, exegetical/homiletical works), Philostorgius, and the Oracula Sibyllina. Not found on IA: Eusebius Werke 2,1-2 (the Kirchengeschichte text itself - only the Teil 3 apparatus volume appears), Epiphanius Bd 3 (Holl 1933), Didymus der Blinde, Methodius.

Licence: not stated

Access: No series-level dump. Per item, everything is a plain file under https://archive.org/download/<id>/ : <id>_djvu.txt (plain text, 0.8-3.3 MB), <id>.pdf (46-59 MB for a 500-page volume), <id>_djvu.xml (per-word text + bbox + x-confidence), <id>_hocr.html (19-40 MB), <id>_chocr.html.gz (gzip hOCR), page/n<leaf>.jpg (JPEG page image, 200). HTTP Range is supported on the text files (206), so a 40 KB sample from any offset costs nothing. · sample HTTP 200

Not reachable: HathiTrust: https://babel.hathitrust.org/cgi/ls?... -> 403 and https://hdl.handle.net/2027/mdp.39015025315527 -> 403 (Cloudflare challenge) from this host on 2026-11-14. Canonical alternative for the same GCS volumes: the Internet Archive copies above; HathiTrust adds nothing that would justify unblocking for this corpus.; Google Books API: https://www.googleapis.com/books/v1/volumes?q=%22griechischen+christlichen+Schriftsteller%22 -> 429 (quota, no key) - same as recorded in the register; the Google-digitised copies that are usable are mirrored on IA as bub_gb_*.; e-rara: the SRU endpoint now answers 200 with a browser-verification interstitial ("Verifying your browser") instead of SRU XML for both dc.title="christlichen Schriftsteller" and dc.title="Patrologia", so it could not be counted in this pass. Canonical alternative: e-rara's IIIF manifests still resolve (https://www.e-rara.ch/i3f/v20/<VLID>/manifest -> 200 per the register) and the Swiss corpus is thin for patristics anyway.; MDZ (digitale-sammlungen.de): no query API - the site is a JS SPA, and a bogus IIIF id returns 404 while real ids need to come from the BSB catalogue; not countable from here. Its patristic holdings are already covered by the hathi-google register entry.; ÖNB: IIIF v3 manifest https://api.onb.ac.at/iiif/presentation/v3/manifest/1008BEB0 -> 200 (endpoint alive), but discovery (onb.digital) was 504 for the register and was not retried in this pass; no GCS volumes are expected there.; Internet Archive item didymusderblinde00jaha_0 (Didymus der Blinde): /metadata returns no 'metadata' block and the download 503s - the item is withdrawn or restricted; Didymus is therefore a genuine gap on IA, not a discovery failure.; Internet Archive item eusebiuswerkeach0008karl (Eusebius Werke 8: Praeparatio Evangelica): _djvu.txt -> 401 (lending copy, no public text layer).

1) Enumerate by identifier pattern, not by series title: 'identifier:diegriechischenc*', 'identifier:origeneswerke*', 'identifier:clemensalexandri*', 'identifier:eusebiuswerke*', 'title:("Hippolytus Werke")' against https://archive.org/advancedsearch.php?output=json&rows=100. The series-title query…

khazarzar-pg-migne: Migne Patrologia Graeca split into per-work Greek PDFs (Library of Ruslan Khazarzar; catalogue at patristica.net)

Measured by crawling the directory index on 2026-11-14: 140 author folders holding 1,545 files, all of them PDFs (one, 'John Chrysostom_PG 47-64/In Genesim sermo 3', is a 15-page PDF that merely lacks the .pdf extension); a seventh file at the directory root, 'A_Kleida Patrologias-1.doc', is a key to the series. Folder names reference 143 of the 161 numbered PG volumes; there is no folder for PG 78-79, 107, 123-126, 128-130, 137-138, 141, 145-147, 152 or 161. Largest author folders (file counts are per-work splits): John Chrysostom PG 47-64 = 403; Athanasius PG 25-28 = 119; Gregory of Nyssa PG 44-46 = 83; Origen PG 11-17 = 81; Cyril of Alexandria PG 68-77 = 74; Basil PG 29-32 = 69; Gregory of Nazianzus PG 35-38 = 68; John of Damascus PG 94-96 = 52; Eusebius PG 19-24 = 44; Epiphanius PG 41-43 = 43; Theodoret PG 80-84 = 27. Includes whole works that are hard to find elsewhere as clean text: Photius Bibliotheca (593 pp), Cyril De adoratione (341 pp), Zonaras Epitome historiarum (438 pp), Didymus Commentarii in Ecclesiasten (64 pp). Why this matters here: the project's own gap dossiers for these works record the Greek e-text as an Internet Archive volume scan and, for Amphilochius, state outright 'No open Greek e-text was located' (/var/www/naxi/web/data/christian/gaps.json, entry greek-amphilochius-iambi-seleucum). This host is the first source found that gives these works as per-work, extractable Greek text rather than a 600-page scan, so it converts several 'scan-only' gaps into ingestable text.

Licence: not stated

Access: Per-work HTTP GET of small PDFs (88 KB - 11.3 MB each); HTTP Range is honoured (byte-range request returned 206 with Accept-Ranges: bytes). 140 folder listings + 1,545 file links were crawled on 2026-11-14 with 140/140 HTTP 200. robots.txt is served but empty (HTTP 200, 0 bytes), so there is no stated crawl restriction; the host posts no rate-limit or terms page for this directory. · sample HTTP 206

1) Fetch http://khazarzar.skeptik.net/pgm/PG_Migne/ (200) and take the 140 folder links (skip '?C=N;O=D' sort links, which are not files). 2) For each folder fetch its index (200); keep links ending .pdf AND test the extension-less ones - 'John Chrysostom_PG 47-64/In Genesim sermo 3' is a real 15-p…

latin-patristic: Latin patristic and medieval Latin full text at Corpus Corporum (University of Zurich) — Patrologia Latina tomus 1-221, CSEL (Scriptores Ecclesiastici), Acta Sanctorum, Aquinas

Measured from the live corpus index (navigate.php?load=cps, HTTP 200, 18,681 bytes, 2026-11-14): 30 corpora, 10,049 texts. Latin patristic/medieval cores: Patrologia Latina 5,277 texts / 5,204 works / 1,528 authors / 85,525,824 words — corpus description: "Patrologiae cursus completus. Series Latina / accurante J.-P. Migne. Parisiis. Tomus 1 (1844) - tomus 221 (1864)", so the whole of PL is transcribed, far beyond the 85 volumes in the Leipzig upstream repo (verified: Bernard, Sermones in Cantica canticorum, TEI header says 'Patrologia Latina, vol. 183'). Scriptores Ecclesiastici (corpus 19) = CSEL: 272 texts / 5,845,390 words — description: "Corpus Scriptorum Ecclesiasticorum Latinorum, Vienna: Österreichische Akademie der Wissenschaften 1864–. This corpus contains all volumes not covered by copyright and digitised by OpenGreekAndLatin project ... They were added from 2016 to 2021." Acta Sanctorum (corpus 28): 66 texts = the 66 AASS volumes (Ianuarius 1 ... December), 52,680,193 words, e.g. 'Acta Sanctorum, Ianuarius 1' = 1,219,589 words. Thomas Aquinas (corpus 3): 77 texts / 8,095,449 words — GATED (see license). Other corpora that matter for Latin Christianity: Grammatici Latini 103 texts/1,071,297 words (incl. Augustine's Ars breuiata and Donatus/Priscian material); Latinitas antiqua 197/5,258,738; Antiquitas Posterior 189/2,979,900 (late antique, incl. digilibLT-type texts); Mirabile Digital Library 406/6,260,466; Monumenta Germaniae Historica 406/6,643,810; Latinità Italiana del Medioevo 348/4,371,143; Hagiographica 10/276,081; Theologica 12/214,022; Poetica 681/2,858,922; Libri sacri 7/2,451,449 (bible versions); Graeca miscellanea 775/10,842,709 (Greek texts with Latin translations). Author-level counts for the patristic core, from navigate.php?load=38: Augustinus Hipponensis 140 works/5,145,629 words; Hieronymus 104/2,309,087; Ambrosius 45/1,103,857; Tertullianus 29/342,701; Cyprianus 18/152,231; Lactantius 12/174,321; Hilarius Pictaviensis 16/441,263; Isidorus Hispalensis 23/453,459; Beda 51/1,560,135; Gregorius I 26/1,564,033; Bernardus Claraevallensis 72/790,922; Boetius 41/956,667; Cassiodorus 11/655,287; Petrus Lombardus 4/1,174,636; Rabanus Maurus 57/2,814,524.

Licence: Mixed, and the mix must be resolved per text. (1) SITE STATEMENT (rendered on the home page, read 2026-11-14): "The data we transcribed ourselves (including some editions by Phili…

Access: No corpus zip ('type=archive-xml-and-data' returns the 3,148-byte SPA shell). Bulk = iterate texts: navigate down from cps -> author -> work -> text idno, then GET download.php?idno=<text idno>&type=file-xml. Expect 15-30 s per navigate call and one navigate per work, so a full PL ingest (~5,277 texts) is ~10k requests; cache aggressively. Files are POS-annotated TEI (EpiDoc/CTS for CSEL), 1.9 MB zip -> 22.7 MB XML for De civitate Dei. · sample HTTP 200

Not reachable: documentacatholicaomnia.eu — https:// fails certificate verification ('unable to verify the first certificate'; leaf CN=documentacatholicaomnia.eu, no intermediate served). http:// redirects to https. With curl -k the index returns 200 (17,473 B) and its sections are reachable (00_De_Sacrae_Scripturae_Libris.html, 07_De_Sanctis..., 01_De_Ecclesiae_Magisterio.html), but nothing could be verified end-to-end as a machine-readable corpus in this sweep, and a TLS-broken host is a bad ingestion dependency.; la.wikisource.org MediaWiki API — first two calls returned HTTP 429 ('You are making too many requests to the API'). It recovers with a descriptive User-Agent and >=2 s between calls (then 200): statistics 84,626 pages / 21,751 articles / 95,865,080 article words; 207 subpages 'Patrologia Latina/1' ... '/207'.; mlat.uzh.ch single-page search — the Sphinx search UI cannot be driven by URL (state is in JS), so only browse/API paths are scriptable; the full-text index itself is reachable only through the browser app.

WORKING RECIPE (all verified 2026-11-14, HTTP 200, from outside the UZH network). 1) Enumerate: GET https://mlat.uzh.ch/php_modules/navigate.php?load=cps&group_by= -> XML with every corpus, its idno/nr/texts_count/words_count. 2) Per corpus: load=<corpus idno> gives the author list with idnos and w…

loc-sinai-arabic: Library of Congress — Manuscripts in St. Catherine's Monastery, Mount Sinai: Arabic microfilm series 5014 (Christian Arabic manuscripts of Sinai)

308 items tagged language=arabic inside the LOC St Catherine's collection (312 items sit under the partof series 'Microfilm 5014: Arabic'; the collection as a whole is 1,691 digitised items, of which 1,078 are 'Microfilm 5010: Greek' and 156 Syriac). Each Arabic item is a complete microfilm roll digitised as a page-image sequence: e.g. 'Arabic Manuscripts 309. Patristica' = 237 folios → 241 IIIF files; 'Arabic Manuscripts 308. Mimars' = 424 folios → 433 files; 'Arabic Manuscripts 321. Barlaam and Youassef' = 310 folios → 314 files. Nine items carry 'canons' in the checklist title (Arabic Manuscripts 390, 391, 392, 393, 445, 600, 677, 316 and one more) — an Arabic canonical corpus that no other source in the register covers. Titles are the LOC checklist headings only ('Commentary (Genesis)', 'Mimars', 'Patristica', 'Epistles of St. Paul') — no incipit, author or edition identification. Content is predominantly Christian Arabic (Gospel commentaries, patristic florilegia, homiliaries, lectionaries, canon law, hagiography) because it is the Arabic collection of a Christian monastery, but the series also holds some Islamic material — do not assume every item is Christian.

Licence: not stated

Access: none (no bulk download, no zip, no OCR text). Page images are fetched one at a time from the IIIF image service: https://tile.loc.gov/image-services/iiif/service:amed:amedmonastery:<item-id>:<seq4>/full/full/0/default.jpg — <seq4> is a zero-padded 4-digit sequence (0005 is the first framed image in the examples checked). · sample HTTP 200

Not reachable: https://www.loc.gov/item/00279383990-ms/ (HTML) — HTTP 403 from this host (Cloudflare interstitial); use the ?fo=json endpoint instead.; https://www.loc.gov/item/00279383990-ms/manifest.json — HTTP 403 Cloudflare ('Just a moment...'); no IIIF manifest retrievable, only the image-service tiles.; Biblioteca Apostolica Vaticana's own Sinai digitisation and the St Catherine's monastery in-house library are not public; the LOC microfilm is the open surrogate.

1) Discover: GET the collection JSON with fa=language:arabic&fo=json&c=100 and page with &sp=2..4 (308 items). 2) Per item: GET https://www.loc.gov/item/<id>/?fo=json — the HTML detail pages (/item/<id>/, /resource/<id>/) return HTTP 403 to this host through Cloudflare, the JSON returns 200; the II…

myriobiblos: Myriobiblos / Μυριόβιβλος, the e-text library of the Church of Greece (myriobiblos.gr)

Small. 37 distinct /texts/greek/ URLs appear across the five Greek index pages, ~34 of them documents; the 'Patristic texts' page (patrology_paterika_en.html) lists about 20 authors (Andrew of Crete, Athanasius, Basil, Germanus of Constantinople, Gregory Nazianzen, John Chrysostom, Nicholas Cabasilas, Maximus the Confessor, Nilus, Photius, Symeon of Thessaloniki, Symeon the New Theologian, Theodore the Studite). Crucially, most are modern-Greek translations, not the ancient text. Verified individually on 2026-11-14: Akathistos hymn = genuine Byzantine Greek (9,165 Greek codepoints); Gregory Nazianzen Or. 31 'Λόγος λα' περί του Αγίου Πνεύματος' = modern Greek (δημοτική); Nilus 'Κεφάλαια περί Προσευχής' = modern translation by Archim. E. Vittis; Maximus 'Μυσταγωγία' is labelled '(στην δημοτική)'; Symeon the New Theologian Catecheses = Greek original retyped from Sources Chrétiennes; the Liturgy of Chrysostom is split into an introduction plus 4 pages of HTML (chrysostom_liturgy.htm -> chrysostom_liturgy1..4.htm, each 200). English section (texts/english/) holds translations, e.g. basil_spiritu.html, gregory_pentecost.html, photios_mystagogy.html, alleenneil_maximus.html.

Licence: not stated

Access: Direct file GETs only; each text is one HTML page (or a 4-page set for the Liturgy). No archive, no manifest. 37 distinct /texts/greek/ URLs were collected from the five Greek index pages: roughly 34 documents plus index/PDF entries. · sample HTTP 200

Harvest the index pages, not the site search: texts/greek/index_gr.html plus patrology_paterika.html, liturgical.html and akolouthies.html give all 37 /texts/greek/ URLs. Strip tags and the shared navigation block (it repeats the same 30 menu labels on every page); the payload starts after the line…

Claremont Colleges Digital Library — Nag Hammadi Archive (Institute for Antiquity and Christianity, Claremont Graduate University)

3,474 items (API totalResults, 2026-11-14). The collection mixes: (a) photographs of the thirteen Nag Hammadi codices and their leaves — item titles give codex and papyrus page numbers, e.g. ptr 2705 'Codex II, papyrus pages 90 and 51', ptr 1381 'Codex II, opened at pages 50-51', ptr 2580 'Codex II, papyrus pages 32 and 111', ptr 2034 'Codex VI papyri pages 6 and 71', ptr 2045 'Codex IV papyri pages 28 and 57', ptr 1228 'Codex XII papyrus page 33'; (b) excavation and site photographs at Jabal al-Tarif; (c) portraits of scholars and officials (e.g. ptr 727 '1956 International Committee' with Quispel, Puech, Guillaumont, Labib). Keyword 'Codex' returns 2,437 hits (includes description text). Images are high-resolution: ptr 2705 is 7,320 x 6,023 px.

Licence: not stated

Access: https://ccdl.claremont.edu/digital/api/collection/nha/id/<ptr>/download returns the full JPEG2000 (e.g. ptr 2705 = 15,128,490 bytes, 7320x6023). No collection-level zip found; page it at 100 records per request. Do NOT use OAI-PMH: every nha record in oai_dc is marked status="deleted" (datestamp 2012-02-09). · sample HTTP 200

Not reachable: https://brill.com/display/db/COGN — 405 with a JavaScript CAPTCHA ('Human Verification ... you need to verify that you're not a robot'). Coptic Gnostic Library Online is subscription-only; canonical alternative for the Coptic text: the printed CGL volumes (Brill) or the Nag Hammadi facsimile edition.; https://archive.org/details/facsimileedition0002unse — 200 but access-restricted-item=true (lending only; the IA copies of the Brill 'Facsimile edition of the Nag Hammadi Codices', Codex II/III, and of 'Nag Hammadi codex VIII' are inlibrary/printdisabled). Not usable for bulk ingestion.; OAI-PMH https://ccdl.claremont.edu/oai/oai.php?verb=ListRecords&metadataPrefix=oai_dc&set=nha — 200, but all 200 returned records carry status="deleted"; the harvestable metadata route is the CONTENTdm API, not OAI.

1) Enumerate with /digital/api/search/collection/nha/maxRecords/100/start/<n> until totalResults (3,474) is exhausted; each item gives itemId, title, collectionAlias. 2) For each item, fetch /digital/api/singleitem/collection/nha/id/<itemId> to get the field map (title, descri, notes, source, right…

openn: OPenn (University of Pennsylvania Libraries) — Walters Art Museum Ethiopic/Geez codices (16 items) plus Penn Ms. Codex 2338, delivered as static BagIt bags with TEI descriptions and full-resolution TIFF/JPEG

16 Geez items at the Walters Art Museum, all digitised: W.768 Ethiopic Psalter with Canticles (94 images), W.784 Ethiopian prayer book Takla Maryam (180), W.788 Ethiopian scroll with angels and talismans (8), W.829 Anaphora of Mary / Mass book (88), W.835 Gondar Homiliary (298), W.836 Ethiopian Gospels (504), W.838 Ethiopian canon tables (16), W.839 single leaf, Christ's entombment (2), W.840 single leaf, Saint Luke (2), W.844 two Ethiopian prayer books (136), W.845 Ethiopian scroll with the Lion of Judah (14), W.850 Ethiopian Gospels, Gunda Gunde (431), W.927/W.928/W.929 double-sided sensul leaves (2 each), W.931 Mawas'et, hymns for the funeral ritual (230). Image counts were read off each item's 300/ directory listing. Plus University of Pennsylvania Ms. Codex 2338, 'Ethiopian liturgical manuscript', 131 parchment leaves, written in Geez before 1914 (probably 19th c.), 2 columns of 23 lines. OPenn holds no Ethiopic items outside these two repositories: grepping all 55 per-repository contents CSVs for ethiop/geez/abyssin matched only 0002_contents.csv (Penn) and 0020_contents.csv (Walters).

Licence: not stated

Access: Per bag: GET .../data/W.850/300/ (directory listing, 431 .tif names), then fetch each .../300/W850_000200_300.tif (33 MB per page at 3133x3855, measured). Service JPEGs are .../sap/W850_000200_sap.jpg (~1465x1800, ~0.8 MB). Manifest-md5.txt gives the whole file list with checksums for one request. No zip/tar bundle. Penn bag: /Data/0002/mscodex2338/ (data/marc.xml, data/extra/{master,thumb,web}/). · sample HTTP 200

Not reachable: https://manuscripts.thewalters.org/ and https://art.thewalters.org/detail/<id> return HTTP 403 to every request from this host (the museum's own viewer is Cloudflare-fronted); OPenn is the working mirror for the same images.; No IIIF on openn.library.upenn.edu: https://openn.library.upenn.edu/iiif/2/... returns 404, and the host serves a static Apache tree only.; https://openn.library.upenn.edu/Data/0020/Data/WaltersManuscripts/W850/data/W.850/master/W850_000010_778.tif returns 404; master file names are listed in manifest-md5.txt with the '_600' suffix for some items, so read the manifest instead of guessing (e.g. data/W.850/master/W850_000415_600.tif).; OPenn's Apache answers rapid serial requests with 404s (no Retry-After); >=2 s spacing plus one retry is required.; https://gundagunde.digital.utsc.utoronto.ca (219 MSS from the same monastery as W.850) is behind an Anubis proof-of-work gate and was not bypassed; see the vHMML reading room as the canonical alternative.

1) Enumerate from the HTML index, not the CSV: GET https://openn.library.upenn.edu/Data/0020/01_ACCESS_WALTERS_MANUSCRIPTS.html and extract <li> blocks matching 'Geez' (16 items, each with shelf_mark, title, date range, place, and ./Data/WaltersManuscripts/html/<W>/ + description.html links). 2) Pe…

orthodoxethos: Orthodox Ethos (orthodoxethos.com), English-language catechetical site of the 'Orthodox Ethos' initiative, built on Wix

680 blog posts (counted from blog-posts-sitemap.xml on 2026-11-14), all in English, essay/book-excerpt format. Content is overwhelmingly modern: St. Paisios of Mount Athos, St. Porphyrios, St. Ignaty Brianchaninov, St. Raphael of Brooklyn, Elder Athanasios of Meteora, with occasional patristic excerpts (St. Symeon the New Theologian, St. Maximos the Confessor) quoted inside modern essays. No ancient Greek text, no critical editions, no Greek-script pages. It is a catechetical blog, not an e-text library.

Licence: not stated

Access: 680 posts, one HTML page each. No archive. · sample HTTP 200

Only useful for translation-status clues on modern ascetic literature, and even then weakly, because posts quote rather than edit. If enumerating: fetch sitemap.xml, then blog-posts-sitemap.xml, take <loc> values (680), and fetch each post; strip the Wix chrome (the page is a single React payload,…

paths: PAThs — Archaeological Atlas of Coptic Literature (ERC Advanced Grant 2015, Sapienza Università di Roma): Coptic literary works by Clavis Coptica (CC) number, their authors, manuscripts, colophons and find-places, with the Clavis/CPG/BHG/BHO/BHL/CAVT/CANT/CAe cross-references

From the project's own description (src/components/Intro/works/works.md, read 2026-11-14): 'This section of the PAThs database contains the literary works preserved in the Coptic language, even including the fragmentary ones, both original compositions and translations of Greek works. They are recorded and identified by the Clavis Coptica (CC) entries composed of a 4-digit number preceded by, for example, CC 0599 … Currently, about 1050 works are recorded, but their number is constantly growing.' Tables exposed by the API: manuscripts, places, works, authors, titles, colophons, persons, collections. Each CC entry is linked, where available, to CPG, BHG, BHO, BHL, CAVT, CANT and Clavis Aethiopica (CAe), which is exactly the cross-walk the project needs between the Coptic, Greek and Ethiopic corpora. Counts of manuscripts/places were not measurable without the API key.

Licence: not stated

Access: No dump. The atlas repository (github.com/paths-erc/atlas) holds only application code and 151 GeoJSON files under public/places-geojson/ (placeography), not the works/manuscripts tables. · sample HTTP 401

Not reachable: https://bdus.lad-sapienza.it/paths/api/… — HTTP 401 {"status":"error","code":"unauthenticated"} without the project-issued api_key; the atlas UI itself loads but renders from that same API

Blocked on the API key. Once obtained: GET /paths/api/records/works?page=N (30 records/page), then /paths/api/record/works/{id} for the full record with related manuscripts and claves; join to the project's Coptic corpus on the CC number, and to Greek on CPG/BHG. The licence (CC BY-NC-SA 4.0) must…

Patrologia Orientalis — Coptic fascicles (editions with Latin/French/English translation), Internet Archive copies

Patrologia Orientalis vols. 1-49 (Turnhout/Paris, 1904-), an ongoing series of eastern Christian texts with translation. 73 IA items were returned for title:(patrologia orientalis); they overlap heavily (the same volume appears as patrologiaorienNNpari, patrologiaorientNNpariuoft, patrologia-orientalis-volume-N and Google/bwb copies), covering vols. 1-25 densely plus scattered later volumes. Coptic-bearing fascicles verified 2026-11-14 by full-text-inside search: vol. 2/2 'Les Apocryphes coptes: Les évangiles des douze apôtres et de saint Barthélemy' (Revillout, 1907; 20 hits for 'apocryphes coptes', pp. 133-); vol. 9/2 'Acta Pilati' (Revillout, 1913; 53 hits, p. 69-); vol. 11/3 'Histoire d'Isaac, patriarche jacobite d'Alexandrie ... texte copte' (Porcher, 1915; p. 305-); vol. 14/2 'The Life of Abba John Kamé, Coptic text edited and translated from Cod. Vat. Copt. LX' (M. H. Davis, 1920; p. 9-); vol. 18/2 'Le livre de Job: version copte bohairique' (Porcher, 1924; 173 hits, p. 9- and 219-). Also Coptic-related but in copyright: vol. 24/2 (Burmester, Holy Week Lectionary, 1933) and vol. 35/4 (Saint Claude d'Antioche, 1970). Adjacent, not Coptic text: the Arabic Synaxaire arabe-jacobite (vols. 1, 11, 16, 17, 20) and the Arabic History of the Patriarchs (vols. 1, 5, 10). Overlap note for the merge: two sibling lane files (sources-more-arabic-christian.json, sources-more-documenta-catholica-omnia.json) also cite Patrologia Orientalis — for its Arabic/Syriac/Geʿez fascicles. Deduplicate by fascicle (vol./part), not by series: the Arabic Synaxaire and the History of the Patriarchs are theirs, the five Coptic fascicles listed in new_texts are this dossier's.

Licence: not stated

Access: Per-item PDF, JP2 zip, *_djvu.txt (Range requests supported). No single PO collection item; harvest the 73 identifiers and dedupe by volume. · sample HTTP 200

Not reachable: https://archive.org/metadata/patrologia-orientalis-29 — 200 but the upload ('Patrologia Orientalis 20, 24, 26-49', 278 files) has no date, no rights and includes in-copyright 1970s-80s volumes; use it only for volume 24 awareness, never as a PD source.; https://archive.org/details/patrologiae-cursus-completus (Migne) — not this source; Migne has no Coptic (see the existing ia-migne dossier).

1) Harvest identifiers with the advancedsearch query above, then group by the metadata 'volume' field (the numeric suffixes patrologiaorienNNpari / patrologiaorientNNpariuoft are reliable). 2) To find a specific Coptic work, do NOT grep the djvu.txt (Coptic OCR is garbage). Use the full-text-inside…

pemptousia: Pemptousia (Πεμπτουσία), Greek Orthodox portal of the Holy Monastery of St. Maximos the Greek (pemptousia.gr) and its English site pemptousia.com

A magazine/portal, not a text library: the homepage (331,282 B) carries 30+ thematic rubrics (Αγιολογία, Ερμηνεία Αγίας Γραφής, Θεολογία, Άγιον Όρος, Συναξάρι...) and the index is article-by-article. The articles *about* the Fathers are modern Greek commentary: the sample homily page (215,938 B) opens 'Όταν μελετήσεις κανείς αυτήν την ομιλία του Μ. Βασιλείου, γεννιούνται οι εξής σκέψεις' and continues in demotic Greek, with a modern-Greek 'Σύντομος Βίος' appended - that is a meditation on Basil's Εἰς τὸ «Πρόσεχε σεαυτῷ», not the Greek text of the homily. No systematic ancient-language corpus was found on the reachable pages.

Licence: not stated

Access: None usable: no archive, no manifest, one HTML article per page. The English twin site https://pemptousia.com/ returns 200 (63,142 B). · sample HTTP 200

Not reachable: https://www.pemptousia.gr/vivliothiki/ -> 403 (the 'ΨΗΦΙΑΚΗ ΒΙΒΛΙΟΘΗΚΗ' menu item has no reachable public listing).

Treat Pemptousia as a *reception* source, not a text source: it is useful for locating modern-Greek translations and for seeing which patristic works are being quoted in Greek Orthodox catechesis, and useless as a witness of the ancient text. If it must be crawled at all: search with ?s=<term>, tak…

pinakes: Pinakes | Πίνακες — Textes et manuscrits grecs (IRHT-CNRS), the Greek Index Project database of every Greek text to the end of the 16th c. contained in catalogued manuscripts (papyri excluded)

Scope statement from /presentation.html: every Greek text from the origins to the end of the 16th century found in printed manuscript catalogues, manuscripts of any date, papyri excluded; 1,300 libraries. Size at launch (same page): "Au moment de sa mise en ligne en 2008, la base Pinakes compte 200 000 fiches ; elle recense les témoins de la tradition manuscrite de 13 000 œuvres, soit près de 40 000 manuscrits conservés dans 1300 bibliothèques du monde entier." No current total is published; measured today, the author authority list alone holds 4,953 name entries over 3,034 author ids. Each work notice carries a witness table (country, city, repository, fonds, shelfmark, folios, century, date, recension, BHG number, incipit/explicit, bibliography) with a per-witness 'Numéro diktyon' and Arca id; each cote notice carries the contents (author/work/CPG/BHG/TLG) and, when digitised, a 'Visualiseur' link (verified: cote 67696 = Vat. gr. 1065 → https://digi.vatlib.it/view/MSS_Vat.gr.1065). Cross-references: TLG, CPG, BHG, BHO, BHL; BHG numbers link to Clavis Clavium (e.g. https://clavis.brepols.net/clacla/OA/Link.aspx?clavis=bhg&number=420m, which itself 302s to a login).

Licence: not stated

Access: No dump, OAI or SPARQL found. Author authority list can be had in one request (above); work and witness data must be scraped notice by notice (one GET per work + one per witness cote). · sample HTTP 200

Not reachable: https://clavis.brepols.net/clacla/ — HTTP 302 to a login (Clavis Clavium subscription); the BHG links from Pinakes land there; https://pinakes.irht.cnrs.fr/api/oeuvres/get/informations/{id} — HTTP 200 but an IRHT 'Service temporairement indisponible' guard page, not JSON; the notice HTML is the working route

1) Pull the author authority list once: GET /api/auteurs/autocompletion/letter_first → dict of {n: {id, displayName}}; dedupe by id (4,953 entries collapse to 3,034 ids). 2) For a chosen author, GET /notices/auteur/{id}/ and parse the <tr> rows: the work title is the second cell text and the link i…

sbb-berlin: Staatsbibliothek zu Berlin - Preußischer Kulturbesitz, 'Christlich-orientalische Handschriften' (Christian Oriental manuscripts) digitised set — 51 Ethiopic (gez) records, OAI-PMH + METS/MODS with leaf-level contents, Public Domain Mark 1.0

51 Ethiopic (gez) records in the digitised Christian-Oriental set (743 records total; the rest are Arabic 258, Syriac 212/168, Coptic 137, Armenian 107, Nubian 14, Greek 12, Georgian 10). Note the discrepancy: Beta Masaheft's digitised-manuscripts page (https://betamasaheft.eu/availableImages.html) says 86 Berlin manuscripts 'as of Aug. 2026', while dc:language=gez in this OAI set yields 51 — the extra items are presumably untagged or outside this set, so 51 is the verifiable floor. Shelfmarks and leaf counts read from the MODS of 14 records: Ms. or. quart. 1014 (150 ll.), 1166 (70), 283 (77), 1001 (81), 1165 (64), 370 (98), Ms. or. oct. 234 (81), 1299 (20), Petermann II Nachtrag 34 (97), 35 (101), 45 (141), Hs. or. 9834 (2 illustrations), Hs. or. 9846 a (2 ll.), 9846 b (1 l.). Contents include the Gädlä Wälättä Ṗeṭros with its miracles and salām hymns, a Taʾamrä Māryām with 85 miracles, a Mäṣḥafä Ḥasab (calendar computation), a 20-item homiliary for the feasts of the church year, a Proverbs/Ecclesiastes/Wisdom/Isaiah/Daniel codex, a Sirach/Daniel/Esra codex including the Apocalypse of Esra, and the Sane Golgotha Marian prayer. Overlap warning: Beta Masaheft already holds Berlin catalogue records (e.g. https://betamasaheft.eu/api/iiif/BerlinMsOrQuart1014/manifest returns 400 'no manifest', i.e. the record exists but is not imaged there), so SBB adds the digitised-image layer, the leaf-level German content descriptions and the PDM statement, not new shelfmarks.

Licence: not stated

Access: Metadata bulk is easy (OAI-METS, one request per manuscript, ~0.1–0.6 MB). Images are not: the IIIF Image API host https://content.staatsbibliothek-berlin.de is reachable but answers every tested path (/iiif/image/<PPN>/<file>/info.json, /iiif/presentation/<PPN>/manifest, /iiif/<PPN>/manifest, with and without the 'PPN' prefix, v2/2 variants) with the same 221-byte plain-text self-description ('This is the NGCS-Chula, an image server that implements the IIIF Image API Level 2'), and its https://content.staatsbibliothek-berlin.de/robots.txt says 'User-agent: * / Disallow: /'. The METS fileGrp only carries the library-internal paths (file:///goobi/tiff001/sbb/PPN1737534649/00000001.tif), not a fetchable URL. Images are therefore viewer-only (browser, one leaf at a time). · sample HTTP 200

Not reachable: Image delivery: https://content.staatsbibliothek-berlin.de/iiif/... returns 200 with the same 221-byte text ('This is the NGCS-Chula, an image server that implements the IIIF Image API Level 2...') for every path tried (/iiif/image/<PPN>/<file>/info.json, /iiif/presentation/<PPN>/manifest, /iiif/<PPN>/manifest, /iiif/image/2/ and /v2/ variants, with and without the PPN prefix), and https://content.staatsbibliothek-berlin.de/robots.txt is 'User-agent: * / Disallow: /'. https://ngcs-beta.staatsbibliothek-berlin.de/docs (the server's own documentation link) does not resolve from this host (curl exit 6/000).; The search and viewer front end is a Meteor SPA: https://digital.staatsbibliothek-berlin.de/werkansicht?PPN=... and /suche?... both return the identical 4,126-byte application shell at HTTP 200, so no record data is retrievable without a browser.; The 1830s DFG-viewer mirror http://digital.staatsbibliothek-berlin.de/dms/werkansicht/?PPN=... 301-redirects into the same SPA shell.; Beta Masaheft overlap: BerlinMsOrQuart1014, BerlinMsOrQuart1165, BerlinPetermannII34 and BerlinMsOrQuart370 all answer 400 ('no manifest') on https://betamasaheft.eu/api/iiif/<id>/manifest, confirming the catalogue records already exist there without images — do not double-count them as new shelfmarks.

1) Metadata: paginate https://oai.sbb.berlin/?verb=ListRecords&metadataPrefix=oai_dc&set=christlich.orientalische.handschriften via resumptionToken (50/page, 743 total), keep dc:language=gez (51). The OAI identifier is 'oai:digital.staatsbibliothek-berlin.de:<PPN>' and the same PPN appears in dc:id…

st-takla: St-Takla.org — Coptic Orthodox Arabic/English library (Arabic 'Kotob Keptya' book index + complete English Synaxarion)

Two distinct things. (1) A complete English translation of the Copto-Arabic Synaxarion, presented as 'Complete Coptic Synaxarium (Full Coptic Orthodox Calendar: Daily Synaxarion)' — one HTML page per commemoration under /books/en/church/synaxarium/<MM>-<month>/<day>-<month>-<slug>.html, e.g. /08-bermodah/30-baramouda-mark.html (HTTP 200, 42,059 bytes). Day pages are reachable only through the liturgical-calendar navigation, so a crawl must walk the month directories rather than the index. (2) The Arabic 'Kotob Keptya' library: a few hundred Coptic Orthodox books in Arabic (394 links on the index page), overwhelmingly modern pastoral, apologetic and catechetical works by Pope Shenouda III, Anba Bishoy, Fr Tadros Malaty and similar — not critical editions and not ancient texts. No Arabic-language Synaxarion and no Arabic History of the Patriarchs was located at the obvious paths (/books/ar/church/synaxarium/… and /Full-Free-Coptic-Books/Coptic-Synaxarium-or-Synaxarion_Arabic/… both HTTP 404). A third section, 'تاريخ البطاركة (باباوات الأسكندرية)' at /Saints/Coptic-Synaxarium-Orthodox-Saints-Biography-00-Coptic-Orthodox-Popes/Coptic-Popes-History_000-index_.html (HTTP 200, 213,580 bytes, 733 links, with century and name sub-indexes), is a modern per-pope biography series; it was NOT verified to reproduce the Arabic text of Severus ibn al-Muqaffaʿ's History of the Patriarchs, so it is not registered as a text source.

Licence: not stated

Access: none; plain HTML crawl. Pages are served as windows-1256 — decode with cp1256, not UTF-8, or every Arabic string is mojibake. · sample HTTP 200

Not reachable: https://st-takla.org/books/ar/church/synaxarium/ — HTTP 404: no Arabic-language Synaxarion under this path.; https://st-takla.org/Full-Free-Coptic-Books/Coptic-Synaxarium-or-Synaxarion_Arabic/Arabic-Synexarium-or-Synexarion-index.html — HTTP 404.; https://st-takla.org/Copyright.html , /terms.html , /about.html — all HTTP 404, so no licence text could be quoted.

1) Decode as cp1256 on every request (Content-Type: text/html; charset=windows-1256) — UTF-8 decoding throws. 2) The Synaxarion is only reachable by walking month directories: /books/en/church/synaxarium/<NN>-<monthname>/<DD>-<monthname>-<slug>.html; the index page is a liturgical-calendar widget t…

Syriac Christian textual corpora: Digital Syriac Corpus (Oxford-BYU) + Syriaca.org + The Digital Wright (British Library Syriac MSS) + DigiVatLib Vat.sir + Dukhrana Peshitta + HMML Syriac collections

Syriac (and Syriac-script Garshuni) Christian literature -- the class of material completely absent from the 20 registered sources, all of which are Greek, Coptic or Ethiopic. MEASURED 2026-11-14: (A) Digital Syriac Corpus, 632 TEI texts / 2,262,025 words of transcribed Syriac / 610 distinct work titles / 46 distinct author strings. Composition by author (files): Ephrem the Syrian 112, anonymous 104, Isaac of Nineveh 82, Narsai 81, Isho'yahb III of Adiabene 52, Jacob of Serugh 39, Dionysius bar Salibi 31, Aphrahat 24, Paul the Apostle 13, Philoxenos of Mabbug 13, Eusebius of Caesarea 10, Rabbula of Edessa 10, Ps.-Clement of Rome 9, Timothy II 7, Bar Hebraeus 5, John the Apostle 4, plus 30 single-work authors (Bardaisan, Joshua the Stylite, Simeon of Beth Arsham, John Rufus, Moshe bar Kepha, John the Solitary, Thomas of Edessa, Balai, Abraham of Nathpar, Isaac of Antioch, Nestorius, Titus of Bostra, the evangelists and apostles). Complete Peshitta New Testament in 27 files (Gospels 14,056+11,620+19,884+15,893 words; Acts 19,141; Romans 7,357 ... Revelation 7,751). Also Acts of Judas Thomas (24,293 words), Ps.-Clementine Recognitions 1-4 and Homilies 10-14, Eusebius HE books 1-5+7-10, Eusebius Theophania (53,057 words), Chronicle of Ps.-Joshua the Stylite (15,420), Life of Peter the Iberian (12,676), Plerophoriae of John Rufus (20,297), Dionysius bar Salibi's literal commentary on most of the OT (Genesis 26,344 ... Malachi 910), Moshe bar Kepha On the Soul (24,799), Philoxenos Discourses 1-13, Jacob of Serugh memre (On Joseph 1-10 plus ~60 liturgical memre), Isho'yahb III Letters 1-52. English: 4 of 632 files carry a <div type="translation"> (686-689, letters to Papa bar Aggai; 313-925 English words each) -- translation coverage is effectively zero. (B) Syriaca.org (sitemap count, 2026-11-14): 46,068 URLs = CBSS bibliography 29,702 + persons 8,031 + places 3,491 + works 2,830 + taxonomy 1,593 + John of Ephesus 420; TEI at /{type}/{id}.xml, CC BY 3.0. This is a catalogue/authority layer (titles, incipits, editions, translation bibliography, manuscript links), not editions -- but it is the join key for everything else (Digital Syriac Corpus titles carry ref="http://syriaca.org/work/NNNN"). (C) Digital Wright / British Library Syriac MSS: 580 TEI manuscript records in git, 1,150 addressable /ms/{id} URLs (records + composite parts), each with Wright's catalogue text, date, contents mapped to Syriaca.org work ids, and links to the archive.org page of Wright's printed catalogue. (D) DigiVatLib Vat.sir: 373 digitised Syriac manuscripts (shelfmarks Vat.sir.1-659), 146 openly viewable / 227 'limited'; IIIF manifests answer 200 for both classes with BAV copyright attribution. (E) HMML Reading Room: 8,426 of 264,037 catalogued objects have Syriac in their language fields (Iraq 3,299, Turkey 2,138, India 1,067, Lebanon 784, Syria 690, Jerusalem 333); 8,395 marked viewableOnline; 2,264 mention Garshuni in the record; leading Syriac collections CFMM 621 (Church of the Forty Martyrs, Mardin), OLM 410, CCM 389, DFM 373 (Dominican Friars of Mosul), SMMJ 333 (St Mark's Monastery, Jerusalem), CPB 323, MGMT 302, ASOM 290, SOAA 258 (Syriac Orthodox Archdiocese of Aleppo), MBM 248. No transcriptions anywhere in HMML -- metadata + images only. (F) Dukhrana Peshitta tool: BFBS/UBS Peshitta NT verse-by-verse plus Old Syriac Curetonian Gospels and the Sinaitic Palimpsest, interleaved with Etheridge/Murdock/Lamsa English; no bulk export.

Licence: not stated

Access: Git (the only real bulk route): `git clone --depth 1 https://github.com/srophe/syriac-corpus.git` = 632 TEI files, 13.4 MB packed / 48 MB on disk, all CC BY 4.0. `git clone https://github.com/srophe/syriaca-data.git` (792 MB) = Syriaca.org works/persons/places/CBSS TEI. `git clone https://github.com/srophe/britishLibrary-data.git` (132 MB) = 580 Digital Wright manuscript TEI records. HMML: https://www.vhmml.org/datasets/vhmml_rr_fulldata.zip (50,784,411 B -> 757,760,065 B JSON, 264,037 objects, Last-Modified 2026-11-09). DigiVatLib: one manifest per shelfmark, no archive. · sample HTTP 200

Not reachable: e-Ktabe: no such host exists from this machine -- ektabe.com, e-ktabe.com, ektabe.org, ektabe.net, e-ktabe.org, ektabe.io, ktabe.org, ktabe.net all NXDOMAIN (curl exit 6); ktabe.com returns HTTP 200 but is a 114-byte parked-domain lander that redirects to /lander (HTTP 307). Canonical alternative for the same need (open Syriac text editions): the Digital Syriac Corpus, https://syriaccorpus.org/ (632 TEI texts, CC BY 4.0).; 'Dux': no Syriac resource by that name could be located. dux.syriaca.org, dux.bethmardutho.org and dux.syriac.uk are all NXDOMAIN. The nearest working resource (and probably what was meant) is Dukhrana Biblical Research, https://dukhrana.com/peshitta/ (HTTP 200): Peshitta NT + Old Syriac Curetonian + Sinaitic Palimpsest verse-by-verse with Etheridge/Murdock/Lamsa English, no bulk export and no site licence (only https://dukhrana.com/peshitta/copyrights/peshitta_nl_copyright.txt, which is CC BY-NC-ND 3.0 and covers the Dutch translation alone).; British Library IIIF: https://bl.digirati.io/iiif/ark:/81055/vdc_100054718159.0x0 -> HTTP 403 with an S3 AccessDenied XML body ('Access Denied'), with and without a Referer header; https://iiif.bl.uk/ -> HTTP 403; https://imagesonline.bl.uk/ -> HTTP 403; https://www.bl.uk/manuscripts/ -> 404 and /manuscripts/Default.aspx -> 404 (the old Digitised Manuscripts paths are retired). The new BL page https://www.bl.uk/collection/digitised-manuscripts-archives?ref=add_ms_12138 answers 200 and embeds manifest URLs on bl.digirati.io, but those manifests cannot be fetched from this host. Canonical alternative: the Digital Wright catalogue (https://bl.syriac.uk/ms/{n}, TEI at .xml) for the catalogue text, plus the Stanford DASH excerpts (below) for pre-1150 Syriac folios by permission.; https://sinaipalimpsests.org/ -> TLS failure 'no alternative certificate subject name matches target host name' (curl exit 60). Canonical alternative: the UCLA Sinai Library, https://sinaimanuscripts.library.ucla.edu/ (HTTP 200) and https://sinai.library.ucla.edu (HTTP 200).; https://syriaca.org/api/works/list/json?perpage=1, /api/persons/list/json?perpage=1 and /sparql?... all return HTTP 404 with an XML error body -- Syriaca.org exposes no Srophe list API and no SPARQL endpoint. Use the sitemaps (https://syriaca.org/sitemap-1.xml, -2.xml; 46,068 URLs) for counting and .xml per record for data.; https://syriaca.org/mss/index.html, /mss/1, /ms/1 -> HTTP 404: Syriac.org does not host a manuscript subsite; the manuscript layer lives at https://bl.syriac.uk and in HMML/PEMM. Likewise https://syriaca.org/bhse/1.xml -> 404 (BHSE pages are HTML only; /bhse/index.html is 200).; https://syriaccorpus.org/{id}.xml -> HTTP 403 (XML export disabled on the public app); same for https://bl.syriac.uk/about -> 403. Fetch TEI from raw.githubusercontent.com/srophe/{syriac-corpus,britishLibrary-data} instead.; https://www.vhmml.org/readingRoom/search?keywords=syriac -> HTTP 403 ('Turnstile validation failed'); POST /dataPortal/search and /dataPortal/export/* are equally unusable -- the nightly zip is the only bulk route (as recorded in the existing `hmml` entry).; https://hugoye.bethmardutho.org/volumeindex.html -> HTTP 403 and /volume/29 -> 0 article links (its issue page is /current-issue.html instead); the working index is https://hugoye.bethmardutho.org/volumes.html (29 volume pages, 681 article links across volumes 1-28). Hugoye is included here only as a secondary CC BY 4.0 corpus of Syriac text editions in article form, not as a primary harvest.

Recommended pipeline, in this order. (1) Digital Syriac Corpus first: `git clone --depth 1 https://github.com/srophe/syriac-corpus.git`; enumerate data/tei/*.xml (632 files, numeric ids 1-692 with gaps); work key = the file id, displayed URL = https://syriaccorpus.org/{id}. (2) PITFALL -- the Syria…

trismegistos: Trismegistos (KU Leuven / Universitätsbibliothek Leipzig) — the cross-database portal for texts from the ancient world, in practice the authority layer for Greek, Coptic, Demotic and Arabic papyri, ostraca and inscriptions (TM numbers, dates, places, archives, TM Places, TM People)

Not measurable from this host: 403 on the root and on every path tried. What Trismegistos is known to provide (from the project's own descriptions elsewhere, not verified today): TM Text identifiers for papyrological, ostraca, wooden-tablet, inscription and related texts with dates, provenance and bibliography; TM Places; TM People; TM Archives; TM Authors (including Christian authors with CPG/BHG cross-references). Counts must be taken from the live site or the dataservices documentation once reachable.

Licence: not stated

Access: Unknown — the download/API pages could not be fetched. · sample HTTP 403

Not reachable: https://www.trismegistos.org/ — HTTP 403 'Security Check' page loading /ab_bot-check.js on every path (tested 2026-11-14); needs a JavaScript-capable client or an allow-listed network; https://www.trismegistos.org/dataservices/ — same 403; https://papyri.info/search?q=Anastasius — HTTP 200 but an Anubis v1.27.0 proof-of-work challenge page ('Making sure you're not a bot!'), so the intended canonical alternative for papyrus texts is also JS-gated here; CANONICAL ALTERNATIVES that do work: the DDbDP/partner EpiDoc corpus as a git checkout — https://api.github.com/repos/papyri/idp.data returned HTTP 200 (repo 'papyri/idp.data', master, ~3.0 GB, no licence field declared by the API), and the HGV/APIS data in the same organisation; provider IIIF for the physical objects (e.g. Chester Beatty via Europeana, see sources-more-europeana.json)

Cannot be written honestly until the host answers. When it does: TM numbers are the join key to papyri.info (DDbDP/HGV/APIS ids) and to the Leuven databases, so the useful ingestion is (a) TM id → partner id mapping, (b) TM Places/TM People authority ids for the gazetteer and person records, (c) th…

ub-frankfurt-msorient: Universitätsbibliothek Frankfurt am Main, Digitale Sammlungen 'msorient' (Orientalische und Slawische Handschriften) — 35 digitised Ethiopic (Geez) codices, including the Hiob Ludolf manuscripts, with per-page JPEG and per-volume PDF

35 Ethiopic (gez) codices digitised and catalogued, out of 60 records in the 'msorient' OAI set (the rest are Syriac, Arabic, Hebrew and Slavonic). Contents fall into: 10 Psalters (Ms. or. 2, 9, 19, 20, 23, 24, 25, 26, 28, 30); three copies of the 'Harfe des Lobpreises und Zither der Psalmodie und Leier der Lobpreisung' (Ms. or. 22, 27, 29); Octateuch/Pentateuch and prophets (Ms. or. 10 'Biblische Bücher des Alten Testaments', 616 pp; Ms. or. 11 Jeremiah, 143 pp; Ms. or. 12 Ezekiel, 134 pp; Ms. Ff. H. Ludolf II 21 'Pentateuchus', 508 pp, 1684); canon law (Ms. or. 14 'Gesetzbuch der Könige', 334 pp); chronicles (Ms. or. 38 'Chronik von Abessinien bis auf das Jahr 7301 (1809)', 559 pp; Ms. or. 39 'Große Chronik des Lik Atkum', 449 pp; Ms. or. 40 'Kleine Chronik des Lik Atkum', 308 pp; Ms. or. 41 world history with Axumite decrees, 1736); theology and history (Ms. or. 133, 134 'Allgemeine Welt- und Kirchengeschichte' 422 pp, Ms. or. 13); prophecy (Ms. or. 16 'Buch der weisen Propheten', 258 pp); prayer and hymnody (Ms. or. 15, 17 'Gebete der heiligen Jungfrau Maria' 290 pp, 31, 32, 33 'Salām an Takla Hāymānot' 34 pp, 34 'Gebet Marias auf Golgota' 100 pp); Ludolf's own working copies (Ms. Ff. H. Ludolf II 21 and II 22 'Collectanea Aethiopica'). Page totals above are read from each METS extent or from the count of structMap page divs.

Licence: not stated

Access: No IIIF (https://sammlungen.ub.uni-frankfurt.de/msorient/iiif/<urn>/manifest and .../iiif/presentation/<urn>/manifest both 404). Bulk = per record: GetRecord mets, then the fileSec URLs: whole-volume PDF at https://sammlungen.ub.uni-frankfurt.de/download/pdf/<id> and page images at https://sammlungen.ub.uni-frankfurt.de/download/webcache/<size>/<imageId> where <size> is 304, 1200 or 2000 (200 = up to 2000 px on the long side; 3000 and 4000 return 404, so 2000 px is the ceiling). Image ids come from the METS 'DEFAULT' fileGrp — do not construct them. · sample HTTP 200

Not reachable: The HTML catalogue and facet search are bot-gated: https://sammlungen.ub.uni-frankfurt.de/msorient/nav/index/all?facets=language%3D%22gez%22 and /msorient/urn/<urn> return 200 with a Visual Library 'Verifying your browser / Preparing verification… Please enable JavaScript and cookies to continue' interstitial and no record data.; No IIIF: /msorient/iiif/<urn>/manifest, /msorient/iiif/presentation/<urn>/manifest and /iiif/<id>/manifest all return 404.; Image resolution ceiling: /download/webcache/2000/<id> is the largest derivative (200, ~2,000 px); 3000 and 4000 return 404 with the same 4,559-byte error page.

1) Harvest metadata: GET https://sammlungen.ub.uni-frankfurt.de/msorient/oai/?verb=ListRecords&metadataPrefix=oai_dc&set=msorient, follow resumptionToken (6 pages for 60 records), keep records whose dc:language is 'gez' (35). Keep both identifiers: the dc:identifier 'system:<id>' is the shelf-level…

armenian-georgian-editions: Armenian and Georgian text editions, second sweep — TITUS Old Georgian corpus directories not itemised in the first pass (apocrypha, liturgies, hagiography, Gelati school, inscriptions, Kartlis Cxovreba vol. II), the Scriptores Iberici and Armenian volumes inside the Internet Archive CSCO mirror, the four Armenian fascicles of Patrologia Orientalis, the Robert Bedrosian Armenian library on Internet Archive (scanned Grabar editions of P'awstos Buzand, Ghazar P'arpets'i, Zenob, Orbelean, Narekats'i, Kanonagirk' Hayots', Liakatar Vark', the Abgar/Addai dossier), and pre-1931 Armenian imprints scanned by Google

Four layers, all HTTP-verified on 2026-11-14. (a) TITUS Old Georgian: the first pass itemised 16 works; the corpus list actually names 221 works in 182 corpus directories and I probed 30 previously-unlisted framesets — all 30 returned 200 — of which 10 are entered here as new texts with page counts measured by walking NNN.htm until 404: Balavariani (21,267 B p.1, 652 Georgian codepoints, 40+ pages), Protoevangelium Jacobi (23,892 B, 1,003 cp, 21 pages), Acta Pilati (11,833 B, 396 cp, 17 pages), Adam and Eve (563,053 B, 28,182 cp, 1 page holding two versions), Caverna Thesaurorum S-30 (775,539 B, 40,306 cp, 7 pages), Kartlis Cxovreba vol. II (669,487 B, 39,620 cp, 40+ pages), Arsen Sapareli (8,345 B, 154 cp, 19 pages), Nemesius De natura hominis (369,846 B, 23,212 cp, 40+), Paris Lectionary (397,300 B, 17,798 cp, 40+), Corpus of Georgian Inscriptions (21,803 B, 1,424 cp, 40+ items). Eleven further verified-but-unlisted TITUS Georgian works are noted under the entries below rather than duplicated as rows. The Armenian side of TITUS yields nothing new: all 15 open works (16 text units, counting the Zohrab Bible's OT and NT separately) are already registered and the 8 'Desiderata' works are linkless. (b) CSCO on Internet Archive: 120 volume labels = the 115 main volumes already noted by the register plus 5 extra bindings; the Iberici series is complete for its first six volumes (Iber 1-2 = CSCO 122-123, Iber 3-4 = CSCO 143-144, Iber 5-6 = CSCO 148-149) and Sub. 4 = CSCO 132. Each has image PDF + _djvu.txt + _jp2.zip + hOCR; only CSCO 122-123 and 143-144 also have a '_text.pdf'. (c) Patrologia Orientalis item: 769 files / 69 volume labels; Armenian fascicles present are PO 15/3 (N74, Synaxaire arménien III, mois de Sahmi, Bayan), 18/1 (N86, V, Kalotz), 19/1 (N91, VI, Arats) and 44/2 (Chaîne arménienne) — fascicles I, II and IV are absent. Georgian: only PO 19/5 (N95, Marr's Synaxaire géorgien, already in the register). (d) Robert Bedrosian Library: 3,916 texts, 848 in Armenian script (language=arm), a private scholar's mirror of scanned Grabar editions and his own English versions; primary-text series present include Liakatar vark' (5+ volumes), Vark' srbots' harants' (2), Kanonagirk' Hayots' (2), Manr zhamanakagrut'yunner (4), plus the individual histories used here. (e) Google-scanned Armenian imprints on IA: 668 texts carry language=(Armenian); the pre-1931 Christian subset includes St Cyprian's De unitate (1846), Nerses Shnorhali's letters, Sebeos, Mambrē Vertsanogh's homilies (1894), the 1898 Tōnakarg and the 1901 Kanonagir. (f) Matenadaran: the digital-library page was rebuilt — it no longer lists any PDF (0 matches for .pdf/ftp in the live HTML and in the 2026-08-09 Wayback capture); previously known files still answer (Sebeos.pdf 20,334,328 B, HTTP 200), but no new filename can be recovered from the site without guessing. (g) manuscript.ge (National Centre of Manuscripts) is now reachable — 200, 36,783 B — but the Mokvi Gospels page is prose plus one upload JPG, with no per-page images and no IIIF.

Licence: not stated

Access: Per file only. IA pattern: https://archive.org/download/<identifier>/<url-encoded filename> (spaces %20, commas %2C, parentheses %28%29). TITUS: no archive, no git, Apache directory listings 403; crawl from texte2.htm and then walk NNN.htm pages. Bedrosian items are individually downloadable with no item-level torrent. · sample HTTP 200

Not reachable: https://geomanuscript.ge/ -> curl exit 60: 'SSL: no alternative certificate subject name matches target host name geomanuscript.ge' (HTTP never reached; the National Centre's alternate portal is broken at the TLS layer, whereas https://manuscript.ge/ answers 200).; https://matenadaran.am/en/matenadaran/digital-resources/digital-library/ -> 200 but 0 matches for '.pdf' or 'ftp' in 176,765 B of HTML; the page now carries a Google-Drive link instead of the 115 PDF links recorded by the first pass. The Wayback capture of 2026-08-09 (web.archive.org/web/20260809010549id_/...) is 166,580 B and also has 0 PDF names, so the list cannot be recovered from the archive either. https://www.matenadaran.am/ftp/el_gradaran/ -> 403. Individual known files still live: Sebeos.pdf -> 200, 20,334,328 B.; https://dspace.nplg.gov.ge/simple-search?query=... -> 410 Gone (282 B body) for every query tried, both plain and percent-encoded; the DSpace OAI-PMH endpoint is the only working access (/oai/request Identify 200, ListSets 200).; https://rbedrosian.com/grabar_hist.html -> 404, redirected to https://rbedrosian.com/403.shtml (315 B). The site TITUS links to as 'Classical Armenian Historical Sources' no longer serves the Grabar texts; the same material is now the IA collection robert-bedrosian-library.; https://www.digilib.am/ -> 404 with 0 bytes (TITUS's 'Related content: Digilib' pointer for Armenian is dead).; TITUS Old Armenian 'Desiderata' block (https://titus.uni-frankfurt.de/texte/texte2.htm#aarmen): Anania Sirakac'i, Elishē, Hovhannēs Drasxanakertec'i, Lazar Parpec'i, P'awstos Biwzandac'i, Sebēos, Stepanos Orbelyan and the Sempad Codex are listed with <A NAME> anchors and no HREF — the Armenian corpus offers no further text beyond the 15 works already registered.; https://archive.org/download/Patrologia-Orientalis/PO%2044-2_chaine_armenienne_pierre_text.pdf -> 404 (this fascicle has only the image PDF, 6,824,303 B, and its _djvu.txt); likewise PO synaxaire arménien fascicles I, II and IV are absent from the item, so the Armenian Synaxarion is not complete there.

1) TITUS: parse texte2.htm by splitting on <LI><A NAME=...>; each work's HREF points at a frameset whose FIRST <frame src> is the text page (balavx.htm is the third frame, i.e. the index, and contains no hrefs — do not try to count chapters from it). Page files are NNN.htm from 001 upward; walk 'Ne…

Coptic liturgical service books: free diocese/parish PDF editions (Coptic Orthodox Diocese of the Midlands UK; St. Mary & St. George and St. Antony, Ottawa; St. Bishoy Coptic Orthodox Church, Canada; C.O.P.T. Coptic Orthodox Publication and Translation), Coptic-text HTML editions at coptic.org and st-takla.org, the O'Leary Difnar scans at copticplace.org, and openly downloadable pre-1931 imprints and manuscripts on Internet Archive

Coptic liturgical books as downloadable text, not catalogues. Verified downloadable: 5 Katameros volumes (3,087 pages: Days 640, Lent 708, Sundays 680, Pentecost 390, Holy Pascha 669), each with an English translation alongside Coptic incipits; 3 complete Euchologion editions (347 pp, 560 pp, and the Traditional English volume) which between them carry the Liturgies of St Basil, St Gregory and St Cyril, the 18 fractions, the refill of the chalice and the blessings; the same three liturgies plus the Raising of Incense and the fractions again as Coptic-only HTML at coptic.org; the Arabic Khulaji (1,026 pp), the only volume here that also contains the ordination and burial rites; the deacon's service book (129 pp) and the marriage rite (26 pp); 4 Agpeya editions (134 pp English, 174 pp English C.O.P.T. with EPUB, 315 pp trilingual, French+Arabic PDF) plus an English HTML Agpeya in 8 files at st-takla.org; the Psalmodia in English/Coptic/Arabic (94 pp annual Sundays, 148 pp Kiahk Sundays), in German (443 pp), and as 232 per-hymn HTML pages with English + Coptic transliteration + Arabic side by side; the Theotokia as a 1911 English translation and a 1923 Coptic critical text; the Difnar (Antiphonarium) in 3 scanned parts (127 + 75 + one unmeasured pages); the Liturgy of St Mark in Malan's 1872 English translation and the Coptic ordinary canon of the mass in Swainson 1884; and three 18th-19th c. Psalmodia manuscripts (1791 Coptic+Arabic, 1750 Coptic selections, 1830 Coptic with doxologies). Main gaps: no standalone English ordination or burial rite outside the Arabic Khulaji, no Sahidic liturgical text beyond what the register already holds, and no critical edition (Doresse-Lanne, Hammerschmidt) reachable outside a library.

Licence: not stated

Access: Archive.org: direct GET on https://archive.org/download/<id>/<file> (Range supported; every file listed below answered 206 to a 2 KB range request); each item also carries <id>_djvu.txt, <id>_djvu.xml, <id>_jp2.zip (/full/full/ or /full/!1024,1024/ IIIF-style zooms via https://iiif.archive.org/iiif/<id>$<page>/full/1200,/0/default.jpg - tested on coptictheotokia0000copt$5: 200, image/jpeg, 150 KB). ukmidcopts.org, saintbishoy.ca, stmary-ottawa.org and copticplace.org serve the PDFs as plain GETs with no cookies; st-takla.org needs one GET per hymn (232 for the Psalmodia, 5 files for the Euchologion at coptic.org). · sample HTTP 200

Not reachable: https://archive.org/download/copticliturgyaut0000copt/copticliturgyaut0000copt.pdf - HTTP 401 (Internet Archive lending: 'The Coptic liturgy: authorized by H.H. Abba Kyrillos VI', 1963; access-restricted).; https://archive.org/download/copticliturgyofs0000copt/copticliturgyofs0000copt.pdf - HTTP 401 (lending only; 'The Coptic liturgy of St. Basil (for the faithful)', 1994).; https://archive.org/download/completetranslat00copt/completetranslat00copt.pdf - HTTP 401 (lending only; 'A complete translation of the Coptic Orthodox mass and the Liturgy of St. Basil').; https://archive.org/download/holypaschabeingl0000copt/holypaschabeingl0000copt.pdf - HTTP 401 (lending only; 'The Holy Pascha ... according to the current usage in the Church of Alexandria', 1988).; https://ukmidcopts.org/pdf/ and https://ukmidcopts.org/kotob/ - HTTP 403, directory listings closed; the files must be addressed by name.; https://ukmidcopts.org/downloads - HTTP 404 (no downloadable catalogue page), so the 9 file IDs were resolved one by one.; https://stmary-ottawa.org/wp-content/uploads/2025/04/ - HTTP 403 (listing closed; individual PDFs are 200).; https://www.saintjohnthebaptistcoptic.com/sitemap.xml - HTTP 406 (Mod_Security), so its upload directory could not be enumerated; only the three Katameros files found by search were used.; https://www.suscopts.org/pdf/ - HTTP 404 (the Diocese of the Southern US moved off its static /pdf tree; the readings URLs cited by other sites are dead).; https://digi.vatlib.it/view/MSS_Vat.copt.38 - HTTP 404; the shelfmark behind O'Leary's Theotokia text is not published at that path (DigiVatLib image use is also restricted by its terms).; https://www.copticplace.org/files/coptic/Difnar1.pdf - HTTP 403; the same file is served at /files/Difnar1.pdf (200). Not a rights block, a path block.; https://www.katameros.com/ - HTTP 403 (parked domain); katameros.app/en/ is the live app but returns a Vue SPA shell with no readings in the HTML, so no bulk text.; https://archive.org/details/largman88larg_yahoo_20150414_1558 - 'كتاب الخولاجي' has no downloadable file (mediatype opensource_media, empty file list).; http://khazarzar.skeptik.net and other Greek-only mirrors were not relevant here; no Coptic liturgical text was found outside the sources above.

Order of work: (1) the Euchologion EPUB/PDFs and the Agpia EPUB give the cleanest English with Coptic and Arabic columns; (2) align the coptic.org Coptic-only files to them using the Coptic-font ASCII map (`]`=ⲧ, `C`=ⲥ, `q`=ϣ, `v`=ⲫ, `;`=ⲑ, `,`=ϩ, `x`=ⲭ, `[`=ϣ? check per glyph, `=` mid-word marks a…

Shenoute of Atripe, Besa and the White Monastery - editions outside Coptic Scriptorium (Internet Archive public-domain editions incl. CSCO Sinuthii opera omnia and MIFAO 23; open-access journal editions in TYCHE, KU Leuven Lirias, BIFAO/Persée, Annales islamologiques; Internet Archive lending copies of Kuhn's Besa and Pseudo-Shenoute)

20 concrete Shenoute / Besa / White-Monastery pieces below, drawn from: (1) seven public-domain Internet Archive editions - Amelineau 'Oeuvres de Schenoudi' t.1 (84.5 MB PDF, Coptic + French) and t.2 (55.4 MB), Leipoldt/Crum 'Sinuthii archimandritae vita et opera omnia' Scriptores Coptici ser. 2 tom. IV (31.8 MB, Latin version of 47 numbered works + 3 additamenta) and tom. V (10.2 MB, Coptic text of 'De vita monachorum' I-XXVII etc.), Nau 1900 Syriac Life (3.8 MB), Amelineau 1889 'Vie de Schnoudi' (25.0 MB), Chassinat MIFAO 23 'Le quatrieme livre des Entretiens et Epitres de Shenouti' (25.5 MB, Coptic text of IFAO Copte 1, fol. 175-375); plus one catalogue-only item, Bouriant's MMAF 8.1 'La bibliotheque du Deir-Amba Shenoudi' (1892, 33.4 MB), which has no editable text in it; (2) five open-access journal pieces - Delhez Canon 6 (11 pp), two Du Bourguet BIFAO articles (85-109, 99-142), Ghica AnIsl 35 (143-161), and three TYCHE articles (12, 22, 18 pp) editing P.Vindob. Coptic documents from the White Monastery archive; (3) three borrow-only monographs. Not counted: ~40 typographic/linguistic studies on Shenoute's syntax (Shisha-Halevy) and the DOP White Monastery Federation survey, none of which are texts.

Licence: not stated

Access: https://archive.org/download/<identifier>/<file>.pdf ; per-item file list from the metadata API. TYCHE OJS galley URLs /index.php/tyche/article/download/<article>/<galley>. · sample HTTP 200

Not reachable: https://gallica.bnf.fr/ark:/12148/btv1b10028555f - HTTP 403 (BNF blocks scripted clients; the Copte manuscripts of the White Monastery held at BNF cannot be pulled without a browser session).; https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=249095&dn= - HTTP 500 (Ljubljana repository copy of 'Shenoute of Atripe: Against the Origenists', the 2026 dissertation/translation surfaced by OpenAlex).; https://www.persee.fr/docAsPDF/bifao_0255-0962_1955_num_55_1_2137.pdf - HTTP 403 for all three Persee articles (BIFAO 55, BIFAO 57, AnIsl 35) even with a browser user-agent, Referer and cookies; HTML full text is 200.; https://digi.vatlib.it/search?query=Borg.copt - HTTP 403 (DigiVatLib search endpoint); the detail page https://digi.vatlib.it/mss/detail/Borg.copt.109 is 200 but is JS-rendered and the IIIF manifest path https://digi.vatlib.it/iiif/Borg.copt.109/manifest returns 404, so no verified Borgia Coptic manifest could be pinned. No Vatican IIIF item is listed above rather than guess a shelfmark.; https://www.sinaipalimpsests.org/ - TLS failure (certificate does not match host name), so the Sinai palimpsest angle could not be checked.; https://www.manuscripts.uni-leipzig.de/ and https://papyri.uni-leipzig.de/ and https://www.ub.uni-leipzig.de/ - DNS does not resolve / connection timeout from this host, so the Leipzig White Monastery codices could not be reached.; https://doi.org/10.71684/orchr.v95i.104643 - connection timeout (Oriens Christianus 95 article on the codicological context of the 'Wiener Verzeichnis' of Shenoute's works); retry later, it is the best catalogue of the Vienna Shenoute manuscripts.; https://www.peeters-leuven.be/ - HTTP 200 but CSCO volumes (incl. Emmel, Shenoute's Literary Corpus CSCO 599/600, and the Canon 8 volume by Boud'hors, IFAO 2013) are sale-only; no open text. Listed as blocked rather than as a text source.; Internet Archive lending copies: https://archive.org/details/letterssermonsof0000besa (Besa, CSCO 157/158), https://archive.org/details/pseudoshenouteon0206unse (Pseudo-Shenoute, CSCO 206/207), https://archive.org/details/lifeofshenoute0000besa (Bell 1983), https://archive.org/details/copticmanuscript0022shen (Young 1993, 'Coptic Manuscripts from the White Monastery: Works of Shenute'), https://archive.org/details/tudedelexicograp0000chri (Cherix 1979, 'Le discours en presence de Flavien'), https://archive.org/details/sinuthiiarchiman0096shen_i6y2 , ...0108shen , ...0129shen (Wiesmann CSCO 96/108/129) - all return HTTP 401 on direct download.; https://www.ifao.egnet.net/publications/catalogue/?nif=BIFAO055_art_08.pdf&nv=0 - connection reset; the per-article PDF files advertised on the BIFAO issue pages (BIFAO055_art_08.pdf, BIFAO057_art_xx.pdf) are not on a public path (four guessed paths return 404), so IFAO offprints must be ordered.

Order of work: (1) Delhez Canon 6 - fetch the Lirias PDF, parse the text layer for Coptic (it is genuinely Unicode) and the French translation, and record the MONB.YJ folio sigla as the unit of alignment; (2) the three TYCHE PDFs - CC BY 4.0, small, but pdftotext gives almost no Coptic (the 2009 an…

DigilibLT — Biblioteca digitale di testi latini tardoantichi (Università del Piemonte Orientale / Fondazione Museo del Tesoro del Duomo, Vercelli): TEI-P5 editions of 463 late-antique Latin prose works, secular plus a Christian subset (Hegesippus, Jordanes, Cassiodorus, Isidore, Hilarianus, Polemius Silvius, Theodosius, Ps.-Augustine)

463 works by 166 authors, all late-antique Latin prose (2nd-7th c.). Verified by parsing https://digiliblt.uniupo.it/teidocs/canone?l=400&o=0 (2,462,957 bytes, 463 distinct DLT ids, 166 AUT ids). Genres are filterable by CAT id and include Cristianesimo (CAT000043) and Religione (CAT000033); the bulk is grammar, medicine, gromatics, rhetoric. Christian / Christian-era items verified with full text: Hegesippus (Ps.) Historiae V, Jordanes Romana + Getica, Cassiodorus Institutiones + Variae (partial), Ps.-Augustine Regulae, Augustini Ars Breuiata, Hilarianus De cursu temporum, Polemius Silvius Laterculus, Theodosius De situ terrae sanctae, Isidore Historia Gothorum + Chronica + Differentiae, Symmachus Epistulae, Collatio Alexandri et Dindimi, Commonitorium Palladii, Asclepius. Text is a re-keyed critical edition (Teubner/CSEL/MGH base stated per work) with apparatus omitted; not a scan.

Licence: not stated

Access: Enumerate /teidocs/canone?l=400 (one HTTP request, all 463 works + author + title), then fetch /teidocs/idno/<ID>/format/xml per work. No bulk archive/ZIP offered. · sample HTTP 200

Not reachable: https://digiliblt.it/classic-digiliblt (old interface) - not followed; the new one supersedes it; /teidocs/searches/fulltext?q=... returns 404 - the form parameter name is different (action is /teidocs/list/by_fts)

1) One request gets the whole catalogue: GET https://digiliblt.uniupo.it/teidocs/canone?l=400&o=0 (2.46 MB). Parse the repeating pattern <h2 class='item-title'><a href='/teidocs/author/AUT######'>NAME</a> ... <span class='canone-teidoc-title'>TITLE</span> ... <a class='canone-teidoc-link' href='/te…

Monumenta Germaniae Historica digital (dMGH) — the MGH's own digitisation of its critical editions (Auctores antiquissimi, Scriptores rerum Merovingicarum, Scriptores rerum Germaniae, Libelli de lite, Leges/Capitularia/Concilia, Epistolae, Poetae Latini, Gesta pontificum Romanorum), served as a page-image reader with per-volume table of contents and backed by Bayerische Staatsbibliothek IIIF manifests

The whole MGH edition corpus as page images. Verified volume counts from the series indexes: Auctores antiquissimi 15 vols, Scriptores rerum Merovingicarum 8, Scriptores rerum Germaniae N.S. 27, Libelli de lite 3, Gesta pontificum Romanorum 1, Capitularia 2, Concilia 15, Poetae Latini aevi Carolini 8, Epistolae saec. XIII 3, Leges (folio) 5. Verified manifests with canvas counts (HTTP 200): bsb00000786 Auct. ant. 1,2 Eugippius (56 canvases), bsb00000788 Auct. ant. 3,1 Victor Vitensis (100), bsb00000792 Auct. ant. 5,1 Iordanes (274), bsb00000795 Auct. ant. 6,2 Avitus (454), bsb00000827 Auct. ant. 15 Aldhelm (795), bsb00000782 Liber pontificalis I (439), bsb00000828 Libelli de lite 1 (676), bsb00000750 SS rer. Merov. 3 (694), bsb00000820 Capitularia 1 (474), bsb00000636 Concilia 1 (299), bsb00000637 Concilia 2,1 (480), bsb00000831 Poetae 1 (663), bsb00000833 Poetae 3 (840), bsb00000517 Epistolae saec. XIII 1 (804), bsb00000877 Leges 3 (723), bsb00000879 Leges 5 (461), plus InC volumes bsb00000747 SS rer. Merov. 1,1 (683), bsb00000683 Cosmas (394), bsb00000689 Thietmar (686), bsb00000690 Ruotger (100), bsb00000692 Notker (183), bsb00000694 Berthold/Bernold (655), bsb00066353 Vitae Coelestini V (294). Text is Latin; the editorial apparatus and indices are part of the page images, so no cleaned text exists.

Licence: Per volume, read from the BSB IIIF manifest's 'license' field, and three different regimes occur in the same series — all three verified here: https://creativecommons.org/publicdo…

Access: Per volume: 1 request for the manifest, then one IIIF image request per canvas (274-840 canvases per volume; ~250-450 KB per 1200-px JPEG, or /full/full/0/default.jpg for native 4961x7016). No ZIP/PDF endpoint that works without a browser: download.digitale-sammlungen.de/BOOKS/pdf_download.pl returns an HTML interstitial requiring JavaScript and a ticked 'Rechtehinweis' acknowledgement. · sample HTTP 200

Not reachable: https://download.digitale-sammlungen.de/BOOKS/pdf_download.pl?id=00000792&ersteseite=1&letzteseite=104 - HTTP 200 but an HTML/JS interstitial with a rights acknowledgement; no direct PDF; https://api.digitale-sammlungen.de/mets/bsb00000792/mets.xml - 404 (no METS/ALTO from BSB for these volumes); https://api.digitale-sammlungen.de/ocr/bsb00000792/00005 - 404 (no OCR text endpoint); https://www.dmgh.de/dmgh_search?q=Getarum ; https://www.dmgh.de/search.htm?q=Getarum ; https://www.dmgh.de/de/fs1/search/simple.html?query=Getarum - all 404 (the beta dMGH has no full-text search); https://www.dmgh.de/mgh_auct_ant_5_1/img/full/mgh_auct_ant_5_1_0001.jpg - 404; the reader's image path in book.js is relative and not served at that URL, so images must come from BSB IIIF

1) Resolve a volume: series page -> /<vol>/index.htm (chapter anchors, printed page numbers) and /<vol>/book.js (grep 'add/bsb(\d{8})' for the BSB id, and br.getPageWidth/getPageHeight arrays for page count). 2) Fetch https://api.digitale-sammlungen.de/iiif/presentation/v2/bsb<ID>/manifest (v3 also…

eotc-liturgy: Ethiopian Orthodox Tewahedo Church liturgical service books — the official EOTC 'Faith and Order' PDF library at ethiopianorthodox.org (Qeddase/Missal in 15 parts, Dəggwa, Mäṣḥafä Ziq, Zəmāre wä-Mäwaśəʾət, Mäṣḥafä Gəṣawé, the complete English Missal of Daoud 1959), with their open Internet Archive counterparts (Harden 1928, Baraki 1984, Ṣəlotä Täktəl) and two Aethiopica text studies of the Säʿatat and the Dəggwa

The complete Ethiopic Missal (Qeddase) as the Church prints it: the Ordinary plus all fourteen anaphoras of the modern EOTC Missal, in 15 scanned PDFs (8-55 pp each, 200 dpi 1-bit, image-only) — this is the Ge'ez text itself, which the register previously knew only through Beta Masaheft work records (LIT1101Anapho, LIT1102Anapho, LIT1569, LIT1931Mashaf etc.); plus a 185-page digital English translation of the same Missal (Daoud 1959/1991, full text layer); plus the chant and service-book shelf: Ṣomä Dəggwa (233 pp), Mäṣḥafä Ziq (226 pp), Zəmāre wä-Mäwaśəʾət (246 pp), Mäṣḥafä Gəṣawé (182 pp), and — inventoried but not itemised below — Dəggwa (28 pp, /amharic/yezemametsheft/zematarik/digua.pdf), Zema (17 pp, .../zema.pdf), Mäṣḥafä Məzmur (91 pp, /amharic/holybooks/2metshafe mezmur.pdf), Mädläwä albab (502 pp, /amharic/holybooks/zema metsaheft/medalewe albab.pdf), 'Book of Christian prayer' (146 pp, /amharic/holybooks/yebetekirstian tselot metshaf.pdf), Gädlä Səmaʿətat (133 pp, /amharic/holybooks/gedele semaetat ethiopians.pdf), and the 394-page Amharic monograph ʾAkkʷateta qʷərbān (2024, /amharic/yezemametsheft/kesis meberatu/Akoteta_Qurban.pdf, 551,682 characters of extractable Amharic text). Outside the EOTC site: Harden's 1928 English anaphora collection, Baraki's 1984 bilingual Ordinary + Anaphora of the Apostles, a Ge'ez marriage rite, and two Aethiopica text studies (Säʿatat, Dəggwa zəyāqon). Caveat on translation_status: in the 15 Qeddase rows it records whether a published English version of the work exists (Daoud 1959, Harden 1928, Baraki 1984), not whether the linked scan is translated — the scans are Ge'ez only.

Licence: not stated

Access: wget -r -np -A pdf on the two EOTC directories; the Qeddase set is ~20 MB for all 15 files, but the chant shelf is large (Mäṣḥafä Ziq alone 189 MB, Metsehafe ziq wemezmur 755 MB, Yetekle Akuakuam Zmame 495 MB). Internet Archive bulk via https://archive.org/download/<id>/<file>; no auth needed for Harden, Ṣəlotä Täktəl, Baraki and the items below. · sample HTTP 200

Not reachable: https://archive.org/details/liturgyofethiopi0000unse — Daoud's Cairo 1959 print copy: files flagged private=true (encrypted PDF/EPUB, 'printdisabled'), lending only; use the EOTC PDF instead (verified open, HTTP 200).; https://archive.org/details/thiopischeliturg0000hamm — Hammerschmidt, Äthiopische liturgische Texte der Bodleian Library in Oxford (1960): private=true, inlibrary/printdisabled, no download.; https://digi.vatlib.it/iiif/MSS_Vat.et.18|28|32|44/manifest — all returned HTTP 404 (no manifest at that path for these shelfmarks), and the detail pages carry no IIIF link in the HTML; Vat. et. liturgical codices could not be verified as digitised.; https://eotc-patriarch.org/ — TLS handshake fails ('unable to get local issuer certificate'); no content retrievable.; https://www.tewahedo.org/ — HTTP 522 (origin unreachable).; https://gallica.bnf.fr/SRU?operation=searchRetrieve&... — HTTP 403 to plain requests; the BnF Éthiopien liturgical manuscripts were not enumerable.; https://html.duckduckgo.com/html/ and https://lite.duckduckgo.com/lite/ — HTTP 202 anomaly page; https://www.mojeek.com/search — HTTP 403. No third-party web search was usable, so discovery was done through site indexes, the Internet Archive advancedsearch API and the Aethiopica OJS search.; https://archive.org/advancedsearch.php?q=qeddase — 0 hits: Internet Archive metadata search matches Latin-script titles only, so Ge'ez/Amharic-titled scans are effectively invisible; searching 'ethiopian liturgy', 'ethiopic liturgy', 'geez liturgy', 'senkesar' was the workable route.

1) URL handling: the EOTC paths contain spaces ('kesis meberatu', 'yaredawi metsaheft', 'zema metsaheft', 'Metshafe gitsawe.pdf'); percent-encode them (%20) or wget will 404. All files were fetched with a plain UA header; no cookies, no auth, no rate limiting observed (about 30 requests, ~900 MB, a…

Ethiopic-monastic: Ethiopian monastic manuscript collections and their text editions — Dabra Dammo (Debre Damo), Lake Tana and Tigray holdings in digitised collections (John Rylands Library / Manchester Digital Collections; Princeton Theological Seminary Ethiopian Manuscript Collection), monastic texts edited in Aethiopica, and public-domain Gǝʿǝz editions on the Internet Archive

Five digitised Ethiopic codices at Manchester (John Rylands), 1,880 IIIF canvases in total: MS Ethiopic 4 (Deggwā, 173 ff, 383 canvases, copied under Iyasu II 1730–55), MS 11 (Dǝrsanä Mikāʾel, 152 ff, 292 canvases, 32 Gondar miniatures), MS 18 (Gebra ḥemāmāt, 246 ff, 500 canvases), MS 23 (Enoch, Job, 1–2 Samuel, 1–2 Kings, 67 ff, 147 canvases), MS 27 (Arbāʿtu wangēl, 275 ff, 558 canvases). Several came from the Church of the Saviour of the World at Āmba Maryam (Mäqdäla) and reached England with the 1867–68 expedition; all are described in Strelcyn's 1974 Rylands catalogue. Princeton Theological Seminary holds 12 codices (Ms. Eth 1–12) plus one 1948 letter, i.e. 13 Internet Archive items (two codices are split into -A/-B parts), all with full-codex PDFs, jp2 bundles and OCR, and all under Public Domain Mark; the monastic-relevant ones are the St Michael devotions (Eth 3), the Mary + St George devotional (Eth 9), two prayer books (Eth 6, 7), a magic-scroll codex (Eth 10) and a fables codex (Eth 12); the rest are Psalters and a 1946 New Testament. Aethiopica supplies the only open, peer-reviewed Gǝʿǝz editions found in this domain: a 15th-c. homily on the Archangel Uriel edited from EMML 1835 + EMML 1841 with English translation (21/2018, 87–119); the monastic genealogies of Hoḫʷärwa edited and translated from Jerusalem, National Library Ms. Or. 87 (23/2020, 57–86); the documents of the Gospel of Däbrä Kärbe (Zana, Tigre) — royal letters, the Śərʿatä däbtära and land donations — edited with English translation (9/2006, 64–79); a pre-14th-c. hymnody fragment from Mǝʾǝsar Gwǝḥila, East Tigray, transcribed with a tentative translation (17/2014, 65–77); and a 41-page handlist of the 'Homiliary for the Year' with Gǝʿǝz incipits from a dozen witnesses including Dabra Ṣəyon Abuna Abrəhām, Tigray (28/2025, 53–92). Aethiopica 25/2022 (99–124) reports the digitisation of 30 codices at the monastery of Däbrä Ṣǝyon (Abunä Abrǝham, Gärʿalta, Tigray) and prints a numbered table of them with project sigla YC-01…YC-71 and Beta maṣāḥǝft CAe IDs — the images themselves are held by the St Yared Center (Mekelle) and are not publicly served. On the Internet Archive, one public-domain CSCO Aethiopici volume (CSCO 33–36, 1904: Acta Marqorēwos) is freely downloadable, while the Gadla Abunä Arägawi (Debre Damo's founder) survives openly only as van den Oudenrijn's 1939 French translation, and the Ge'ez Acts of Abba Ǝsṭifānos (CSCO 620, 2006) as an unlicensed third-party scan.

Licence: not stated

Access: Manchester: no archive download; bulk = one IIIF image request per canvas (1,880 canvases for the 5 codices) at the image API, or the 'Download Image' route per page. Princeton Theological Seminary: whole-codex PDFs and jp2.zip are direct-download (e.g. ptsem_ethiopian_mss_2-1.pdf 11,667,963 B; images.zip 661,057,773 B for 2-B), so bulk is one GET per codex. Aethiopica: one GET per article PDF (https://journals.sub.uni-hamburg.de/aethiopica/article/download/{articleId}/{galleyId}); the six PDFs used here total 21,766,959 B (~21.8 MB). Internet Archive CSCO volume: one GET (10,325,535 B). Politeness: Manchester collection page and manifests answered at ~1 req/s with a browser UA; no 429 seen. · sample HTTP 200

Not reachable: https://www.vhmml.org/readingRoom/search?q=ethiopic — HTTP 403 (Cloudflare); HMML Reading Room (the EMML collection, 7,865 objects) is not scriptable, and EMML images are partner-rights regardless of access.; https://gundagunde.hmml.org — curl exit 000, host does not resolve; no Gunda Gunde image portal found.; https://ethiopianmanuscripts.org — curl exit 000, host does not resolve; the Ethiopian Manuscript Imaging Project has no live web presence at that domain.; https://www.ethiosare.uni-hamburg.de/ — curl exit 000, host does not resolve; the Ethio-SPaRe Tigray manuscript database is offline (its records survive inside Beta maṣāḥǝft).; https://cudl.lib.cam.ac.uk/search?q=ethiopic — HTTP 403 (bot block); no Cambridge Ethiopic list obtained.; https://collections.library.yale.edu/catalog?q=ethiopic&format=json — HTTP 202 with a 0-byte body (challenge page), so no Beinecke Ethiopic records.; https://digital.bodleian.ox.ac.uk/collections/ethiopic-manuscripts/ — HTTP 404; Digital Bodleian's search is a Vue SPA whose only machine-readable artefact is a 733-byte OpenSearch stub (https://digital.bodleian.ox.ac.uk/search.xml, HTTP 200) pointing back at the HTML search, so no Ethiopic item list could be retrieved.; https://www.digitale-sammlungen.de/en/search?query=aethiop — HTTP 200 but a 27,467-byte JS shell with no Cod.aethiop. links in the response; BSB IIIF item ids could not be enumerated without a working search API.; https://sinaipalimpsests.org — curl exit 000, host does not resolve.; Internet Archive lending items: vitaesanctorumin0028gadl (Gadla Basalota Mikāʾel + Anorēwos), vitaesanctorumin0000gadl_z8u7 (Gadla Yared + Panṭalēwon), vitaesanctorumin0056unse (Abäkǝrāzun + Takla Häwaryat), vitaesanctorumin0000unse (Walatta Ṗeṭros), actesdemarhakres0000unse (Actes de Märḥa Krǝstos), actesdesamuelded0000unse (Samuel de Dabra Wagag), collectiomonasti0238unse_45 (Collectio monastica), contendingsofapo0001unse/0002unse — all 'access-restricted-item=true', i.e. borrow-only, no downloadable text.

1) Manchester: GET the collection page, then one manifest per codex; canvas image ids follow 'MS-ETHIOPIC-000NN-000-000MM.jp2' and each resolves at https://image.digitalcollections.manchester.ac.uk/iiif/<id>/info.json (200). Fetch with a browser UA; image API v2 permits /full/max/0/default.jpg per…

Internet Archive: Greek canon-law and Byzantine legal collections (Rhalles–Potles Syntagma, Photius/Balsamon Nomocanon tradition, Zachariae's Jus Graeco-Romanum, Zepos' Jus Graecoromanum, Basilika, Hexabiblos, patriarchal acts, Pedalion)

Internet Archive holds, in machine-downloadable files, essentially the whole printed apparatus of Greek canon law: (a) Rhalles–Potles, Σύνταγμα τῶν θείων καὶ ἱερῶν κανόνων, all 6 volumes (Athens 1852–1859) in two independent 6-volume items (SyntagmaTonTheionKaiHieronKanononTonTeHagionKaiPaneuphemonApostolon with djvu.txt + text PDF + EPUB per volume; Vol.2SyntagmaTnTheenKaiHierenKanonn, same 6 volumes, an alternate scan) plus 3 duplicate single-volume scans (syntagmatnthenk00/01/02potlgoog = vols 1, 5, 6; bub_gb_-1lQRyCImeQC = 1852 vol. 1); (b) the Photius Nomocanon tradition in four separate editions — 1561 Morelius, 1615 Justellus (with Balsamon), 1620 Beveridge Synodicon, 1661 Voellus/Justellus (2 vols, 2 IA items) — plus PG 104; (c) Balsamon's council commentary as a 1582 standalone (SBB digark); (d) Byzantine imperial legislation: Zachariae's Jus Graeco-Romanum, 7 vols (Leipzig 1856–1884, one item with all seven, and separate per-volume items for II–VII), Zepos & Zepos Jus Graecoromanum vol. VII (1931: Prochiron Auctum, Psellus, Attaliotes, 26 Chomatianos decisions), Leunclavius' Basilika (1575, Greek+Latin), Harmenopoulos' Hexabiblos (Heimbach 1851), Zachariae's Supplementum editionis Basilicorum; (e) applied canon law: Miklosich–Müller Acta et diplomata Graeca medii aevi, 6 vols (Vienna 1860–1890); (f) the modern-Greek Pedalion (1886) and its 1957 English 'Rudder' (lending-only). Degrees of usability differ sharply (see ingestion_notes): only 2 of the 6 Rhalles–Potles volumes' OCR layers and the Zepos/Zachariae volumes return meaningful Greek, and every 16th–17th-c Greek-type book and every Google-scanned Migne volume has a text layer with zero Greek codepoints.

Licence: not stated

Access: https://archive.org/download/<identifier>/<filename> — per-file GET; no auth, no rate limit observed (26 files fetched in this pass, all 200). Full-item zips exist as <identifier>_<format>.zip for some items but are not needed: the DjVuTXT and *_text.pdf are single files · sample HTTP 200

Not reachable: HathiTrust (babel.hathitrust.org/cgi/ls?q1=rhalles → HTTP 403, Cloudflare 'Just a moment…' interstitial from this host; only the catalog API works: catalog.hathitrust.org/api/volumes/brief/oclc/<n>.json → 200). HT holds the better Rhalles–Potles and Pitra scans but cannot be fetched here.; Pitra, Juris ecclesiastici Graecorum historia et monumenta (Rome 1864–1868, 2 vols): not on Internet Archive (title search 0 hits; creator:(pitra) returns only the Analecta sacra and Spicilegium Solesmense) and not reachable on MDZ. The single most important Greek canon-law edition still missing from the open corpus.; Bayerische Staatsbibliothek MDZ (digitale-sammlungen.de): search returns HTTP 200 but the whole UI is a JavaScript SPA ('this website doesn't work properly without JavaScript enabled'), and api.digitale-sammlungen.de/oai → 404. IIIF manifests do work when the bsb id is known (api.digitale-sammlungen.de/iiif/presentation/v2/bsb10157068/manifest.json → 200), so MDZ is usable only with out-of-band identifiers.; anemi.lib.uoc.gr (University of Crete, the main Greek digital library of 19th-c Greek editions): TLS handshake failure (curl exit 60), and HTTP 403 with -k. Unreachable from this host — would otherwise plausibly hold Rhalles–Potles and the Venetian Pedalion.; The Rudder (Pedalion), Cummings' English translation, 1957 (IA item rudderpedalionof0000agap): access-restricted-item true, collection inlibrary/printdisabled — lending only, no text or images downloadable.; Google Books direct (books.google.com) and e-rara full-text: not fetchable from this host; the usable Google surface remains the bub_gb_* mirror on IA, whose Greek OCR is the failure mode described above.; khazarzar's PG mirror has no PG 104 folder body text usable for the Nomocanon and no 137-138 (already noted in the register); PG 104 therefore comes from the patrologia-volumes container, as page images only.; Archive.org full-text search for Greek ('juris ecclesiastici graecorum', Greek-script title queries) returns noise because the search indexes the broken Latin-model OCR — Greek titles must be found via Latin/transliterated metadata.

Fetch: for each volume take the single file https://archive.org/download/<id>/<file> (DjVuTXT for triage, *_text.pdf for the real text). Every URL written in this dossier was fetched with curl in this pass and returned HTTP 200 (16 text-layer downloads + 27 HEAD/GET checks on PDFs and EPUBs); total…

Internet Archive — J. A. Cramer, Catenae Graecorum Patrum in Novum Testamentum (Oxford: Typographeo Academico, 1844, 8 vols) with Pierre Poussines' Catena Graecorum patrum in evangelium secundum Marcum (Rome: typis Barberinis, 1673)

The complete 8-volume Cramer corpus (Oxford 1844; IA imagecount per volume 546/476/482/552/502/434/622/650 = 4,264 scanned pages) of Greek New Testament catenae, printed as continuous Greek text with the biblical lemma and the excerpted Father named in the margin. Volume map — vols I, V, VI, VII and VIII read from their own title-page OCR, vols II-IV from the IA title field of the parallel bub_gb_ scans (which quote the letter): I = Matthew + Mark; II = Luke + John; III = Acts; IV = Romans; V = 1-2 Corinthians; VI = Galatians, Ephesians, Philippians, Colossians, 1-2 Thessalonians; VII = 1-2 Timothy, Titus, Philemon, Hebrews; VIII = Catholic Epistles (James, 1-2 Peter, 1-3 John, Jude) together with Oecumenius and Arethas on the Apocalypse. The 1673 Poussines volume is an independent, pre-Cramer edition of the same catena and prints it in Latin.

Licence: not stated

Access: Whole item, no auth. The 8 cramuoft volumes: PDF 39,573,566 / 37,062,841 / 31,043,075 / 43,797,766 / 43,190,083 bytes for vols 1, 5, 6, 7, 8 respectively (vols 2-4 fetched earlier in the project), plus one _djvu.txt each (1.2-2.9 MB). Item GRCCRAM_DBS_HS also carries a single merged 8-volume PDF, 143,571,274 bytes. · sample HTTP 200

Not reachable: CCSG (Corpus Christianorum Series Graeca, Brepols) — the modern critical editions that supersede Cramer for several catenae are behind Brepols' paywall; no open PDF route was found and none is claimed here.; Google Books API (https://www.googleapis.com/books/v1/volumes) — HTTP 429, daily quota exhausted ('Quota exceeded ... Queries per day'), so Google Books was not used for any URL in this dossier.

Fetch: /metadata/<identifier> first, then the single .pdf plus the _djvu.txt (both HTTP 200, no auth, no rate limiting observed at ~1 req/s). Parse: the _djvu.txt carries the printed page headers ('TOMUS V.', 'AD EPHES.'), so split a volume by searching for 'CATENA IN ...'/'AD <book>' running heads…

Oecumenius of Trikka / Pseudo-Oecumenius — the Greek commentary chain on the Apocalypse, Acts, the Catholic Epistles and Paul: Migne PG 118–119 as per-work extracts (Khazarzar mirror) and as a full volume (Internet Archive), the 1631 Greek edition, Hoskier's 1928 editio princeps, the 1532 Exegeseis palaiai, and the Arethas / Andreas of Caesarea Apocalypse catenae (Documenta Catholica Omnia)

Migne PG 118–119 (Oecumenius / Pseudo-Oecumenius) in three granularities: per-work Greek PDFs (commentary on the Apocalypse + 14 sets of catena fragments on the Pauline letters), the full PG 118 volume (Oecumenii opera omnia, 1881 reprint, 4.94 MB of OCR text), and an independent 1631 Greek printing of the Hypomnemata (part 2, Greek-only). Two further witnesses to the same chain: the 1532 Exegeseis palaiai (Ps-Oecumenius on Acts, Catholic Epistles and all Paul, with Arethas on the Apocalypse — stated on its title page) and the PG 106 texts of Andreas and Arethas on Revelation, split as per-work PDFs by Migne column range (Arethas: MPG 106 0487–0786).

Licence: not stated

Access: No archive, no OAI, no git. Bulk = one GET per PDF: 15 files from the Oecumenius directory (68 KB – 1.05 MB each), 4 from the Arethas directory (106–164 KB), 2 large Migne per-work PDFs from DCO (40.8 MB Andreas, 41.6 MB Arethas), plus Internet Archive full items (PG 118 volume PDF 55,340,258 B; Hoskier 1928 PDF 14,619,268 B). · sample HTTP 200

Not reachable: CCSG / Brepols critical editions of the Ps-Oecumenian corpus — paywalled; no open PDF.; De Gruyter 'The Pseudo-Oecumenian catena on Ephesians' (2026) — catalogue record only via HOLLIS; the text is not open access.

Order of work: (1) harvest the two Khazarzar directory indexes (plain Apache HTML, one <a href> per PDF) and GET each file; they are small (68 KB–1.05 MB), the Greek is embedded as a text layer from Migne's plates. (2) For the Apocalypse use Hoskier 1928 (IA PDF) rather than PG 118 — Migne's text i…

pg-migne-hagiographica: Migne Patrologia Graeca, per-work Greek PDFs of hagiography and martyr acts (Library of Ruslan Khazarzar, khazarzar.skeptik.net/pgm/PG_Migne/)

Crawled 2026-11-14: 140 author folders / 1,544 PDF links. 81 filenames match hagiographic keywords (Vita|Passio|Martyrium|Laudatio|Encomium|Funebris|Laudes|Laus); of these, 27 are lives, acts, passions or encomia proper and are listed below - the rest are homilies for martyrs' feasts, Marian or dominical homilies, or biographical notices of non-saints (e.g. 'Vita Oppiani', 'Vitae sophistarum'). PG volumes represented in this set: 06 (Justin), 19-24 (Eusebius), 25-28 (Athanasius), 34 & 65 (Palladius), 35-38 (Gregory of Nazianzus), 44-46 (Gregory of Nyssa), 47-64 (John Chrysostom), 80-84 (Theodoret), 85 (Eudocia), 86B (Antonius Hagiographus), 87c (Alexander Monachus), 94-96 (John of Damascus), 140 (Akropolites). Chronological span: 2nd c. (Acta Iustini, Martyrs of Palestine dossier) to 12th-13th c. (Akropolites). Notably absent: 'Symeon Metaphrastes_PG 114-116' holds only Chronicon.pdf and Chronicon breve.pdf (4 rows incl. parent), so the Metaphrastes menologion - the largest single Greek hagiographic corpus - is not served here.

Licence: not stated

Access: Per-file GET; small files (96 KB - 1.7 MB in this set). Range requests are honoured (the sibling dossier measured HTTP 206 with Accept-Ranges: bytes). Ten folder listings and 1,544 file links were crawled on 2026-11-14; all 27 PDFs below were re-fetched with HTTP 200 the same day. HEAD works and returns Last-Modified: Mon, 10 Nov 2008 for the Palladius dialogue, i.e. the mirror has been static since 2008. · sample HTTP 200

Not reachable: https://www.doaks.org/research/byzantine/resources/hagiography-database - HTTP 403 with and without a browser User-Agent (520-byte Apache 'Forbidden' page). The Dumbarton Oaks Hagiography Database (BHG-keyed lives with translation status) would be the ideal cross-check for the translation_status column and is simply unreachable from this host.; https://www.brepolsonline.net/ - HTTP 403, Cloudflare interstitial ('Just a moment...'). No CCSG text is retrievable.; Corpus Christianorum Series Graeca (CCSG): commercial print series, not open. https://www.brepols.net/series/ccsg returns 200 and lists volumes at list price (e.g. 'CCSG 111 (CN 36) | 2027 € 170,00 EXCL.VAT (Hardback)'); the series description says it 'aims to replace Jacques-Paul Migne's Patrologia Graeca, which no longer fulfils the requirements of modern scholarship'. The critical editions of the works below (e.g. Eustratius, Vita Eutychii = CCSG 25) are therefore not ingestible without purchase.; Subsidia hagiographica: commercial. https://bollandistes.org/our-publications/susbidia-hagiographica/ (200) sells the volumes (each entry links to an abebooks listing) and describes the series as 'monographs, critical editions and supplementary reference works ... the famous Bibliothecae hagiographicae, which list all Greek and Latin hagiographic texts up to 1500 A.D.'; the BHG itself is print-only. https://www.brepols.net/series/sh returns 404.; Acta Sanctorum: no open digital text. https://bollandistes.org/our-publications/acta-sanctorum/ (200) states the corpus is '67 in-folio volumes totalizing more than 60,000 pages'. The IA scans of the 1863 Palmé reprint exist (e.g. item actasanctorumquo14boll, metadata 200) but the OCR text layer is unusable: 'actasanctorumquo14boll_djvu.txt' (200, 6,607,190 B) opens with scanner noise ('WAALELILEICIFI] : ^ ILLA A PERS VI or ...') and never resolves into Latin. AASS is also Latin, not Greek, so it is a witness to the Greek acts rather than a Greek text source.; Analecta Bollandiana on Internet Archive: the volumes are present and open (e.g. sim_analecta-bollandiana_1900_19, metadata 200, no access-restriction) but the OCR is too poor to harvest Greek: 'sim_analecta-bollandiana_1900_19_djvu.txt' (200, 612,712 B) yields garbled Latin/Arabic-looking strings ('S.a un culte pour le Lion; ar mire surtout d') and no reliable article headings.; Symeon Metaphrastes, Menologion (PG 114-116): the khazarzar folder 'Symeon%20Metaphrastes_PG%20114-116/' returns 200 with only Chronicon.pdf, Chronicon breve.pdf and an 'extras/' directory - the menologion lives are absent, and no other open per-work Greek text of the Metaphrastes collection was found in this pass.

1) Fetch http://khazarzar.skeptik.net/pgm/PG_Migne/ (200) and keep only folder links ending '/'. 2) Fetch each folder index; filenames are Latin titles ('Vita Antonii.pdf'), folder names carry author + PG volumes. 3) GET the PDF and run 'pdftotext -layout' - no OCR needed: the text layer is Unicode…

Greek Psalm catenae and their constituent commentaries — the Migne per-work PDFs that the Psalter chains are built from (Apollinarius, Diodorus, Eusebius, Origen, Cyril, Athanasius, Gennadius, Ps-Chrysostom, Theodoret), split by work and column range at Documenta Catholica Omnia, plus Daniele Barbaro's 1569 printed catena of fifty psalms

The Greek commentary tradition behind the Psalter catenae, per work, with Migne column ranges in the filenames: Apollinarius of Laodicea's Interpretatio in Psalmos (PG 33, 1309–1538), Diodorus of Tarsus' Fragmenta in Psalmos (PG 33, 1587–1628), Eusebius' Commentaria in Psalmos in five DCO files, Origen's Selecta and Excerpta in Psalmos, Cyril of Alexandria's Expositio in Psalmos in two files, Ps-Chrysostom's Expositio in Psalmos (PG 55, 35–526), Athanasius' Expositiones in Psalmos, Gennadius' fragments (PG 85, 1665–1668), and Theodoret's Interpretatio in Psalmos in three tomes. Add Barbaro's 1569 Latin catena of Psalms 1–50, the only early modern printing that presents a Greek Psalm catena as a chain. These are the works whose excerpts the Byzantine Psalter catenae quote; the catena manuscripts themselves are already indexed in the project (three Biblissima/Handschriftenportal entries for Catena in Psalmos).

Licence: not stated

Access: No dump, no OAI. Bulk = resolve each 04z wrapper page to its ../02g/ or ../20vs/103_migne_gm/ PDF href and GET it: 23 Psalm entries, of which the ones below were fetched with HTTP 200 (5 KB–174 MB each). · sample HTTP 200

Not reachable: Documenta Catholica Omnia has no OAI/JSON API and no bulk archive — only per-file GETs from an HTML index.; CCSG volumes of the Psalm catenae (e.g. the critical editions of the Psalter catena types) — Brepols paywall; not linked.; Hesychius of Jerusalem's Psalter commentary (PG 93) — the Khazarzar PG 93 directory carries only his homilies, and DCO's author index has no Hesychius 'in Psalmos' entry; not found as a per-work PDF.

Fetch: parse the DCO author index once, keep only hrefs matching 04z/z_*Psalmo*; each wrapper page is ~1 KB HTML whose only useful content is the ../02g/... or ../20vs/103_migne_gm/... PDF href. Then GET with curl -g (or percent-encode) — the filenames contain square brackets ('[1]', '[Tomus_1]') w…

syriac-editions: Syriac text editions outside the Digital Syriac Corpus - CSCO Scriptores Syri (Internet Archive mirror), Patrologia Orientalis Syriac fascicles, Bedjan's Acta Martyrum et Sanctorum / Homiliae Selectae Mar-Jacobi Sarugensis / Isaac of Nineveh / Bar Hebraeus, Patrologia Syriaca, Mingana's Narsai, Cureton and Wright, plus two open TEI/TAN corpora (Patristic Text Archive, TAN pseudo-Methodius)

Internet Archive item 'Corpus-Scriptorum-Christianorum-Orientalium': 1,346 files / 187 PDFs, of which 115 are main volumes and 63 are Scriptores Syri (Syr.) volumes, running CSCO 1-155 (Syr 1-74) plus the Syriac subsidia; the Syriac coverage includes the historical chronicles (Chronica minora 1-6, Chronicle of 1234, Ps-Dionysius/Zuqnin, Zacharias Rhetor, Elias of Nisibis, John of Ephesus), the Severus of Antioch corpus (Contra impium grammaticum, Philalethes, letters), Philoxenus, Babai, Timothy I, Jacob of Edessa, Jacob of Sarug's letters, Theodore bar Koni, Išo'yahb III, Dionysius bar Salibi, Theodore of Mopsuestia on John, Cyril on Luke (Syriac), Ephrem (Hymni de fide, In Genesim/Exodum), Išo'dad of Merv. Internet Archive item 'Patrologia-Orientalis': 119 PDFs / 64 main fascicle files, of which ~19 carry Syriac texts (Severus hymns, letters and cathedral homilies; Barhadbešabba; Philoxenus; Narsai; Zacharias' Life of Severus; Syriac martyrologies; John of Ephesus' Lives; the Life of Timothy of Kakhushta). Bedjan on IA: Acta Martyrum et Sanctorum 6 vols, Homiliae Selectae Mar-Jacobi Sarugensis 5 vols (1905-1910), Homiliae S. Isaaci 1 vol (1903), Chronicon Syriacum (1890). Patrologia Syriaca: 3 tomes (Aphrahat Dem. 1-10, 1894; Dem. 11-23, 1907; Liber Graduum, Kmosko 1926). Mingana's Narsai 2 vols (1905). Also reachable but not listed as texts here: Cureton's Eusebius 'History of the Martyrs in Palestine' (1861, ia: historyofmartyrs00euse), Athanasius' Festal Letters (1848, ia: festallettersat00curegoog), John of Ephesus HE part 3 (1853, ia: thirdpartofeccle00john), Lagarde Analecta Syriaca (1858, ia: PLagardiiAnalectaSyriaca) and Libri Veteris Testamenti Apocryphi Syriace (1861, ia: libriveteristes00lagagoog), Budge's Book of the Bee (1886, ia: bookofbee00solo), the Palestinian Syriac Lectionary (1899, ia: palestiniansyria00lagauoft). TEI/TAN: PTA pta0001.pta073 (2 files, Syriac witnesses B and L of one homily); TAN pseudo-Methodius Syriac (1 file, 20,512 Syriac codepoints).

Licence: not stated

Access: Per file only: https://archive.org/download/<identifier>/<url-encoded filename>. Each CSCO volume has an image PDF, a smaller '_text.pdf', a '_jp2.zip' page-image bundle, '_djvu.xml', '_djvu.txt', hOCR and (for most) EPUB. The CSCO item sets 'noarchivetorrent: true', so there is no item-level torrent or bulk tar. · sample HTTP 200

Not reachable: https://syri.ac/ -> HTTP 403 (Cloudflare bot wall) on every attempt on 2026-11-14; the aggregator's link index could not be read, so any Syriac edition it would have surfaced is missing from this sweep.; https://hugoye.bethmardutho.org/volume1num1 -> HTTP 403 while the root returns 200; individual Hugoye volume/article PDFs (which regularly print Syriac text-and-translation editions) could not be enumerated without a browser session.; Gorgias Press 'Gorgias Open Repository': 27 free PDFs under https://www.gorgiaspress.com/images/uploaded/Gorgias%20Open%20Repository/ return HTTP 200 (e.g. 978-1-60724-880-4_QAOK.pdf, which is Kollamparampil's 'Salvation in Christ According to Jacob of Serugh'), but the licence page https://www.gorgiaspress.com/why-open-access-benefits-fees-and-licensing-information 302-redirects to /page-not-found (404) and the PDF's own copyright page reads 'Copyright (c) 2010 by Gorgias Press LLC. All rights reserved ... No part of this publication may be reproduced'. Free access, no usable licence - not ingested.; Peeters (CSCO 540-541 Reinink, Pseudo-Methodius; every Syriac versio volume): https://www.peeters-leuven.be/search_results.php?series=Corpus+Scriptorum+Christianorum+Orientalium returns HTTP 200 but is a JS commercial catalogue with no PDF/API endpoint; the only open copy found for Reinink's edition is the TAN transcription listed above.; Internet Archive advancedsearch query 'collection:corpus-scriptorum-christianorum-orientalium' -> 0 results (HTTP 200, empty docs); the CSCO material is one community item. That item also sets 'noarchivetorrent: true', so there is no item-level torrent or single bulk archive - downloads are per file.; https://archive.org/advancedsearch.php?...creator:(Bedjan) works, but the older Bedjan uploads ia: Actamartyrumetsa01bedjuoft and ia: LibriSuperiorumVariousWorksInSyriac now return empty metadata and no files (dead identifiers).; Cureton's 'History of the Martyrs in Palestine' (ia: historyofmartyrs00euse) and Lagarde's 'Analecta Syriaca' / 'Libri VT Apocryphi Syriace' are reachable but were not opened in this sweep; treated as follow-up, not verified.

1) Discovery: the CSCO Corpus is a single Internet Archive community item, not a collection - GET https://archive.org/metadata/Corpus-Scriptorum-Christianorum-Orientalium returns 1,346 files; filter names matching '^CSCO <num> (Syr <n>) - <editor> <year> - <title> - (t|v)(_text)?\.pdf$'. The 't'/'v…

Theophylact of Ohrid (Theophylactus Bulgariae archiepiscopus, c. 1050–c. 1108) — the Greek text of his Gospel, Acts, Pauline and prophetic commentaries: Patrologia Graeca vols 123–126 as per-volume PDFs from the Internet Archive 'PatrologiaGraeca' set, plus William Gilson Humphry's 1854 Greek edition of the Matthew commentary

Theophylact's complete surviving New Testament commentary in Greek, in the Migne arrangement, which the volumes' own ELENCHUS pages give with column numbers. PG 123 ELENCHUS (normalised from the OCR, which renders the name 'THEOPHILACTUS'): 'THEOPHILACTUS BULGARIE ARCHIEPISCOPUS ... Enarratio in Evangelium S. Matthaei. 143 / — in Evangelium S. Marci. 481 / — in Evangelium S. Lucae. 683 / — in Evangelium S. Joannis. 1127'. PG 124 ELENCHUS: 'Commentarius in Joannis Evangelium. (Continuatio.) col. 9 / Commentarius in omnes divi Pauli Epistolas. 319 / ... in Epistolam ad Romanos. 335 / — in Epistolam 1 ad Corinthios. 563 / ... ad Galatas. 954 / ... ad Philippenses. 1139 / ... I ad Thessalonicenses. 1219'. So: Gospels in PG 123 (John running over into PG 124), Paul from PG 124 into PG 125, Acts in PG 125, minor prophets in PG 126. The same ELENCHUS notes that Theophylact's Psalter commentary was lost and the Venetian editors substituted Euthymius Zigabenus — i.e. there is no Theophylact in Psalmos to look for.

Licence: not stated

Access: One GET per PG volume: Vol. 123 = 103,482,090 B; 124 = 92,496,569 B; 125 = 84,478,788 B; 126 = 84,120,034 B (all HTTP 200, application/pdf), plus a _djvu.txt of 6.3–7.3 MB each. Humphry 1854: 26,772,039 B. · sample HTTP 200

Not reachable: PG 126 scope — the volume's OCR is unreadable and no independent elenchus was retrieved this sweep; only Habakkuk is confirmed.; Google Books API — HTTP 429 quota exhausted, so Google Books was not used.; Brepols / CCSG Theophylact volumes — paywalled.

Fetch: https://archive.org/metadata/PatrologiaGraeca once, then GET each 'Patrologia Graeca Vol. 12N.pdf' (all 200, 84–103 MB). Parse: do NOT try to read the Greek body from the _djvu.txt of PG 123–126 — the Google OCR converted Latin glyphs to Greek lookalikes, so 'THEOPHYLACTUS' appears as 'ΤΗΕΟΡ…

pemm: Princeton Ethiopian, Eritrean, and Egyptian Miracles of Mary (PEMM) catalogue

Ethiopic/Eritrean (and Egyptian/Arabic) Marian miracle literature: the Taʾamməra Maryam corpus plus Life-of-Mary, homilies and apparition stories. MEASURED 2026-11-13 against the live API (perPage=2000 sweep): 1,010 canonical story records (type_of_story: Miracle of Mary 704, Life of Mary 244, Homily 59, null 3); 1,012 manuscripts (976 Geʿez, 35 Arabic, 1 null; 825 with a digital-copy link); 2,473 paintings (2,220 with an IIIF/JPEG image link); sum of per-story total_records = 53,181 story instances (manuscript attestations). English full text present for 821 of 1,010 stories (3.10 M characters, mean 3,777 chars/story) at the default translation_seq, plus 67 stories carrying 2-5 recensions (88 additional sequences) => up to ~909 English text records. 184 stories return the literal placeholder 'An English translation of this story is not yet available.'; 5 return empty. 960 stories have a summary_plot; 990 have a Geʿez canonical_incipit; 666 carry a Beta Masaheft hamburg_id + clavis_id (CAe ...). Other languages offered: Amharic 401, French 181, Arabic 87, Italian 42, Latin 24. The Zenodo 2.0 snapshot is SMALLER and stale: canonical_story CSV 1,055 rows, manuscript CSV 1,022 rows, story_instance CSV 52,049 rows (21,043 typed incipits), but paintings CSV only 84 unique paintings / 162 episode rows — i.e. the archival paintings table is a fragment of the 2,473 live records.

Licence: Split. The archival dataset (Zenodo record 8247393, 'PEMM 2.0', 2023-08-21) is CC BY 4.0 — Zenodo record page: 'Rights License Creative Commons Attribution 4.0 International. The…

Access: https://zenodo.org/records/8247393 (DOI 10.5281/zenodo.8247393, concept DOI 10.5281/zenodo.6909568, version PEMM 2.0). File pattern https://zenodo.org/records/8247393/files/<name>?download=1 works unauthenticated. Files: canonical_story / manuscript / story_instance / translation / paintings / collection / keywords _version2.csv + 7 *_data_dictionary_version2.csv. · sample HTTP 200

Not reachable: https://cms.ethiopicmary.com/items/stories -> 403 FORBIDDEN (Directus public role has no read on stories/manuscripts/paintings/collections); only /items/languages -> 200. The path-style JSON routes (/stories, /manuscripts, /paintings) are the working surface.; https://cms.ethiopicmary.com/robots.txt -> 200 but 'User-agent: * / Disallow: /' — host policy blocks crawling; use Zenodo for bulk.; https://pemm.princeton.edu/stories.json -> 200 but text/html (Next.js catch-all page, not JSON); https://pemm.princeton.edu/api/stories -> 500. No documented API.; https://pemm.princeton.edu/sitemap.xml -> 200 but an HTML page; no sitemap exists.; https://api.github.com/search/repositories?q=... -> 403 API rate limit exceeded for this egress IP; could not confirm or rule out a GitHub mirror of the PEMM code/data. Not needed: Zenodo is authoritative.; https://dataspace.princeton.edu/... -> 401 with an ALTCHA human-verification challenge; no automated access.; No bulk endpoint for story instances or translations on the live API (/translations, /story_instances, /incipits, /keywords all 404) — those tables exist only in the Zenodo CSV snapshot.; Zenodo paintings_version2.csv is materially incomplete (84 unique paintings vs 2,473 live) — treated as a partial export, not a blocked resource.

Recommended path: Zenodo 2.0 CSVs as the licensed, reproducible core; live API for English full text and the post-2023 delta; Beta Masaheft TEI for Geʿez editions via hamburg_id. Steps: (1) mkdir -p /srv/christian-corpus/pemm/{raw/zenodo,raw/api,raw/tei}; curl -L 'https://zenodo.org/records/8247393…

Perseus Digital Library / canonical-greekLit

Ancient Greek literature in CTS/TEI, not a Christian-only corpus. Whole repo (verified against the GitHub git-tree API, recursive=1, truncated=false): 100 text groups, 144 work directories, 1,612 edition/translation XML files + 926 __cts__.xml, 376 MB checkout / 78 MB tar.gz. Christian subset = 16 text groups, 85 works, 114 files (25 MB), 39 Greek files + 75 English files: (1) New Testament, tlg0031 - 27 works (Matthew ... Revelation), 27 Greek files (Westcott-Hort 1882/1892, perseus-grc2) + 27 English files (World English Bible, perseus-eng2); (2) Old Testament / LXX, tlg0527 - 38 works (Genesis ... 2 Ezra, Psalms, 12 prophets etc.), 39 files, ENGLISH ONLY (World English Bible) - no Greek LXX text in this repo; (3) Apostolic Fathers: Didache tlg1311, Epistle of Barnabas tlg1216 (grc+eng), 1-2 Clement tlg1271 (eng x2), Ignatius Epistles tlg1443 (eng), Polycarp to the Philippians tlg1622 (eng), Martyrdom of Polycarp tlg1484 (eng), Shepherd of Hermas tlg1419 (eng), Epistle to Diognetus tlg0646 (eng) - the English is Kirsopp Lake, Apostolic Fathers (Loeb, 1913); (4) Clement of Alexandria tlg0555 - 4 works (Protrepticus, Stromata, Quis dives salvetur, fragment), Greek; (5) Eusebius of Caesarea tlg2018 - Ecclesiastical History, Greek (1.68 MB); (6) Basil of Caesarea tlg2040 - 2 works (Ad adolescentes; Epistulae), Greek; (7) John of Damascus tlg2934 - Vita Barlaam et Joasaph, Greek; (8) Zonaras tlg3135 - Epitome historiarum vol. 1, Greek (opp-grc3). Adjacent but not Christian: Josephus tlg0526, Procopius tlg4029 (2 works, Greek; Christian-era historian, no theological content), Pseudo-Justinus tlg0646 counted above. IMPORTANT: the live Scaife index for these same 16 text groups exposes 128 works / 221 texts (113 perseus-*, 104 1st1K-*, 3 opp-*, 1 INTF-*), i.e. 2.6x this repo, because Scaife merges First1KGreek (Greek LXX, 1st1K-grc1 for tlg0527; 13 Eusebius works; the 1st1K-eng1 translation of Eusebius HE), an INTF-grc1 critical text of Mark, and a Sahidic Coptic Mark (1st1K-cop1). Those extra files are NOT in canonical-greekLit; they are in https://github.com/OpenGreekAndLatin/First1KGreek (local mirror /srv/christian-corpus/f1k: 1,467 textgroups / 1,655 files, likewise CC BY-SA 4.0 with one Coptic-Scriptorium exception). Scaife-wide inventory (https://scaife.perseus.org/library/json/, 6.1 MB, live): 557 text groups, 2,699 works, 3,860 texts across CTS namespaces, of which greekLit = 372 text groups / 1,836 works / 2,817 texts.

Licence: CC BY-SA 4.0 (Creative Commons Attribution-ShareAlike 4.0 International). Repo file license.md opens "## creative commons / # Attribution-ShareAlike 4.0 International"; GitHub API…

Access: git clone https://github.com/PerseusDL/canonical-greekLit.git (verified: git ls-remote succeeds from this host, HEAD dacb04909e43dc47738828d33cc3d2b0635bc54c, 2026-11-13T16:53:05Z). Tarball https://codeload.github.com/PerseusDL/canonical-greekLit/tar.gz/refs/heads/master -> 200, application/x-gzip, 77,623,366 bytes (range requests ignored); zip via https://github.com/PerseusDL/canonical-greekLit/archive/refs/heads/master.zip -> 200. Per-file raw fetch also works: https://raw.githubusercontent.com/PerseusDL/canonical-greekLit/master/<path>. · sample HTTP 200

Not reachable: https://scaife-cts.perseus.org/api/cts?request=GetCapabilities - the documented CTS API endpoint is DEAD: TLS handshake fails with 'tlsv1 unrecognized name' (server presents a cert not matching the host, resolved via sv-cts-prod.eldarion.com = 18.205.222.128/54.161.241.46/52.202.168.65/54.237.133.81), and plain http returns 404 with Heroku's 'No such app' page. Use https://scaife.perseus.org/library/json/ and the /library/.../json/ reader routes instead.; https://scaife.perseus.org/api/cts?request=GetCapabilities - 200 but text/html, 3,544 bytes: it is the single-page-app shell, not a CTS response; the same 3,544-byte shell is returned for any unknown path (e.g. /ctsapi/cts), so a 200 here means nothing.; https://catalog.perseus.org/catalog/urn:cts:greekLit:tlg0031.tlg001.json - 404. The catalog serves HTML only (no JSON, no MODS endpoint found); its data is stale relative to the repo (lists perseus-grc1/eng1/lat1 which no longer exist).; Whole-work download via the reader API is not possible: https://scaife.perseus.org/library/passage/<urn>/text/ without a reference returns 400 {"reason": "URN must contain a reference"}; passages must be enumerated through children[] / refs one at a time.; https://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0155 returns 200 but the page body is 'No document found' - the legacy Hopper id in the catalog does not resolve for this text; Perseus 4 Hopper pages are only reliably reachable through the catalog's HTML record page, and are not a machine-readable path.; Not reachable/not attempted: PDF facsimiles - none exist in this source (nothing blocked, simply absent). No host blocked this machine: GitHub, raw.githubusercontent.com, codeload, scaife.perseus.org, catalog.perseus.org and perseus.tufts.edu all answered 200 from here.

1) git clone the repo (or shallow: git clone --depth 1); do not fetch file-by-file - 1,612 files, 78 MB tarball. 2) Walk data/*/*/*.xml excluding __cts__.xml. Derive the CTS URN from the path (drop .xml), never from file content. 3) Read __cts__.xml at the work level for the title/author/edition-vs…

Coptic manuscript repositories (Vatican, Louvre, BnF, Coptic Museum, CMCL, Coptic Scriptorium)

Vatican Library (DigiVatLib), verified 2026-11-13 from the fond counts in the /mss/ navigation: Vat.copt. 34 digitised manuscripts, Borg.copt. 176, Pap.Vat.copt. 2, plus Borg.eg. 1 and Pap.Hanna. 1 (Coptic-adjacent). That is 212 core Coptic shelfmarks against a stated 31,163 digitised manuscripts overall — i.e. Coptic is ~0.7% of the Vatican's digitised mss. Page counts are large per item: Vat.copt.57 = 573 canvases (a Coptic commentary on Matthew, per the detail-page bibliography citing Lucchesi, 'Deux commentaires coptes sur l'Evangile de Matthieu', Museon 2010), Borg.copt.51 = 1,070 canvases. Image resolution is usable: median 2238x3514 for Vat.copt.57 and 2135x2956 for Borg.copt.51, max 3830px wide. Internet Archive holds the printed-edition complement: 2,534 texts matching (coptic OR copte) and 3,780 with subject:coptic, of which 62 are tagged language:Coptic (Budge's Coptic Psalter, Worrell's Coptic Psalter, Lagarde's Pentateuch, Horner's Northern-dialect NT, Amelineau/Hyvernat martyrdoms, Crum, De Vis). Coptic Scriptorium's OCR repo adds 9 edition families already transcribed (Budge, Amelineau Lausiac History, Amelineau Monks of Egypt, Amelineau Vita Isaac, De Vis Homelies coptes de la Vaticane 1922, Giron Legendes Coptes, Hyvernat and Balestieri Martyrdoms, Lacau Apocrypha, Crum) = 7,774 tracked files, 8.2 GB, including 40+ OCR4All models and a general 'coptic-all' model. CoptOT (coptot.manuscriptroom.com, 200) is a Liferay portal with a Manuscript Catalog and diplomatic transcriptions of the Sahidic OT, but serves no images of its own — it points outward to repositories. CMCL (www.cmcl.it, 200) is the Corpus dei Manoscritti Copti Letterari under Tito Orlandi; its corpus sits behind a login-gated 'CIM server', so only the front matter is public.

Licence: not stated

Access: No bulk archive/dump. Closest to bulk: iterate /mss/ajax/category/<fond> for the shelfmark list, then fetch one manifest per shelfmark, then pull pages individually from the image service. Coptic Scriptorium additionally ships already-OCR'd plain text and TEI-ish XML in a git repo (see ingestion_notes). · sample HTTP 200

Not reachable: https://gallica.bnf.fr/* — 403 from this host on every route tried (iiif manifest, ark page, SRU search API). BnF Coptic manuscripts (Copte collection, e.g. btv1b100915891 cited as bnf78 in Coptic Scriptorium) are therefore unreachable for images here. archivesetmanuscrits.bnf.fr returns 200 for the homepage and for ark:/12148/* records, but its search pages (pageRechercheSimple.html, /recherche) 404, so no Coptic record list could be obtained.; https://copticscriptorium.org/ and https://www.copticscriptorium.org/ — connection timeout (curl 28, 30 s) over both IPv4 and default. The main site is unreachable; the GitHub org (api.github.com/repos/CopticScriptorium, 200) is the working substitute.; https://cmcl.it/ — connection timeout (curl 28, 25 s). Only https://www.cmcl.it/ answers (200, 8.4 KB front page); the Corpus dei Manoscritti Copti Letterari text corpus itself is behind a login-gated 'CIM server' with no public API, so Clavis Patrum Copticorum numbers cannot be resolved online.; https://collections.louvre.fr — 429 on search, and the response body is a bot wall ('WADP - I am not a robot'), i.e. a captcha, not a rate limit you can back off from. IIIF manifest guesses under /iiif/ark:/53355/... returned a 404 site-chrome page, so the Louvre's IIIF prefix could not be determined. No Coptic Louvre object list obtained.; https://www.copticmuseum.gov.eg/ — connection timeout (curl 28, 20 s). The Coptic Museum, Cairo has no reachable online catalogue, so it contributes zero discoverable witnesses.; https://paths.adwmainz.net/ — TLS handshake failure ('tlsv1 alert internal error') on both the bare host and /database. PAThs (An Archaeological Atlas of Coptic Literature), the natural authority for work-to-manuscript linking, could not be queried from this host.; https://www.trismegistos.org/ — 403 with a 'Security Check' interstitial page; no TM numbers resolvable.; https://www.vhmml.org/ — 200 but the Reading Room and Coptic microfilm collections require an account; no anonymous image or metadata API found on the public page.; https://digi.vatlib.it/search, /mss/search, /search/mss — 403 Forbidden (Apache 'You don't have permission to access this resource'). Discovery must go through the fond enumeration instead.; https://coptot.manuscriptroom.com/catalog.json?q=..., /catalog?q=..., /iiif/collection — 404. The portal is Liferay-based and serves no IIIF collection of its own; only the HTML pages (/home, /biblical-books, 200) are usable, and it holds metadata/transcriptions rather than page images.; https://www.sinaipalimpsests.org/ — TLS certificate subject mismatch (curl 60), and https://sinai.library.ucla.edu/ is only the project landing page (200, no IIIF). St Catherine's Coptic palimpsests were not reachable.

Two separate pipelines, because the scanned images and the usable text live in different places. (A) Images from the Vatican: (1) GET https://digi.vatlib.it/mss/ once, regex the left-nav pairs <a href="Vat.copt"><span>Vat.copt.</span><span>(34)</span></a> to get authoritative per-fond counts; the "…

Biblioteca Apostolica Vaticana — DigiVatLib

31,163 digitized items in 83 fonds under /mss/ (counts read from the server-rendered id="fond-<name>" blocks on https://digi.vatlib.it/mss/?ling=en, 2026-11-13; corroborated item-for-item against the /mss/ajax/category/<fond> JSON for 22 fonds). Greek-tradition fonds: Vat.gr 1272, Barb.gr 522, Pal.gr 402, Ott.gr 381, Reg.gr 182, Urb.gr 162, Arch.Cap.S.Pietro 471 (mixed Greek/Latin), Capp.Giulia 104, Reg.gr.Pio.II 42, Borg.gr 8, Pap.Vat.gr 5 (= 3080 strictly .gr, up to ~4192 if Chig 641 — predominantly Greek — and Arch.Cap.S.Pietro are counted in); Chig 641, Ross 1073 and Capp.Sist 407 are cross-classified and mostly Latin, so are excluded from the Greek figure. Coptic: Borg.copt 176 + Vat.copt 34 + Pap.Vat.copt 2 = 212. Ethiopic: Vat.et 48 + Cerulli.et 16 + Borg.et 6 = 70. Adjacent traditions, useful for the Copto-Arabic and Syriac dossiers: Vat.ar 419 + Sbath 50 + Borg.ar 73 = 542 Arabic; Vat.sir 373 + Borg.sir 45 = 418 Syriac; Borg.arm 83 + Vat.arm 43 = 126 Armenian; Vat.iber 2 Georgian; Neofiti 46 + Vat.ebr 606 + Urb.ebr 58 + Borg.ebr 18 Hebrew; Vat.eg 404 (fond not published online) and Borg.eg 1 are the only Egyptological survivals, i.e. Coptic-adjacent material is thin. Page counts: median 209 canvases/item over 35 sampled manifests (mean 269, range 4-1190), so ~644k Greek images, ~32k Coptic, ~14k Ethiopic (estimates, not measured). All items are MANUSCRIPTS/archival objects — DigiVatLib holds no critical editions, so it cannot by itself supply the missing e-text of a work; it supplies witnesses. Script identity verified by OCR, not by the shelfmark prefix: Vat.gr.1205 -> Greek codepoints, Borg.copt.9 -> Coptic (rendered as Greek homoglyphs by a Greek model, 0 Coptic codepoints), Vat.et.1/Borg.et.3 -> Ethiopic fidel, Vat.ar.1 -> Arabic, Vat.sir.1 -> Syriac. Manifest availability is high: 24/24 randomly sampled shelfmarks across Vat.gr, Borg.copt, Vat.et, Vat.sir returned HTTP 200 manifests.

Licence: not stated

Access: No bulk download, no PDF, no DjVu, no OCR dump. Enumeration is cheap: /mss/?ling=en lists all 83 fonds with their item counts in id="fond-<name>" blocks, and /mss/ajax/category/<fond> returns the shelfmark list as JSON. Images must then be pulled one page at a time from the IIIF image service. robots.txt states Crawl-delay: 10, so ~6 requests/minute; at that rate the ~600k Greek page images would take months. Coordinate aggressively or treat this as a per-item, on-demand source rather than a mirror. · sample HTTP 200

Not reachable: https://digi.vatlib.it/mss/search?k_f=0&k_v=coptic (no sm=&rt[]) -> HTTP 403 from Apache/2.4.62; the same query with the full parameter set -> 200. Not a real block, a WAF rule.; https://digi.vatlib.it/mss/index-search?q=<term> -> HTTP 400 (and with sm=&k_f=0&k_v= -> 500; with ?sm=&k_f=0&k_v=coptic -> 200 but zero result rows). The incipit/contents index could not be driven without the site's JS.; https://digi.vatlib.it/pub/digit/MSS_Vat.gr.1205/document.pdf -> HTTP 400; no PDF export exists.; https://digi.vatlib.it/pub/digit/MSS_Vat.gr.1205/Vat.gr.1205_0001.jp2 -> HTTP 400; JP2 masters are not downloadable, only IIPImage JPEG renditions.; https://digi.vatlib.it/iiif/collection.json, /iiif/collection/top, /iiif/collection/Vat.et, /iiif/presentation/MSS_Vat.gr.1205/manifest -> HTTP 404; no IIIF collections.; https://digi.vatlib.it/oai, /oai-pmh, /pub/oai -> 404/404/400; no OAI-PMH.; https://digi.vatlib.it/sitemap.xml -> 404; no sitemap.; https://digi.vatlib.it/mss/Vat.eg -> HTTP 404 (fond not published online); https://digi.vatlib.it/iiif/MSS_Vat.eg.1/manifest.json -> 404 and MSS_Vat.mus.1 -> 404.; https://www.vaticanlibrary.va/home.php?pag=riproduzioni_fotografiche&ling=eng -> 302 to the Italian homepage; the BAV reproduction-rights page is not reachable at the URL the DigiVatLib viewer links to, so the rights text above is quoted from DigiVatLib itself.

Step 1 (enumerate, cheap): parse https://digi.vatlib.it/mss/?ling=en for id="fond-<name>"><div class="item-thumbnail-list-header"><span><strong>N</strong>ITEMS, then GET https://digi.vatlib.it/mss/ajax/category/<fond> (send X-Requested-With: XMLHttpRequest) for a JSON shelfmark list with a 'limited…

Ethiopian manuscript imaging: EMIP + Ethio-SPaRe facsimiles on Beta Masaheft, HMML/vHMML EMML, Sinai Palimpsests Project, and open IIIF of Ethiopic holdings (CUDL, Bodleian, DigiVatLib, LoC)

Beta Masaheft (master branch 2026-11-13T18:14Z, 20,746 TEI records): EMIP 4,410 records (3,042 codices EMIP0xxxx + 1,365 magic-scroll records) of which every numeric EMIP record and every transcribed scroll tested serves a IIIF manifest (18/18 codices 200; 8/8 Transkribus scrolls 200; random untranscribed ms ids mostly return 400 'no manifest available'); Ethio-SPaRe (ES) 1,066 records, 12/16 sampled have IIIF manifests (~75%); EMML 8,190 records, 0/8 sampled have manifests at Beta Masaheft (EMML images exist only inside vHMML); also OxfordBodleian 167, LondonBritishLibrary 587, ParisBNF 340, Berlin 341, VaticanBAV 238, StPetersburg 307, RIE 309, Tanasee 182, Abennat 125 records (mostly metadata-only at Beta Masaheft). Other live IIIF: Cambridge CUDL 13 Ethiopic manuscripts (/iiif/collection/ethiopicmanuscripts), Bodleian 9 (Ethiopian and Eritrean collection), DigiVatLib 72 digitised Ethiopic mss (Vat.et 48, Borg.et 6, Cerulli.et 16, Raineri 2; 40 flagged 'limited'), Library of Congress 6 Ethiopic mss from St Catherine's microfilm 5015 plus 21 from Jerusalem patriarchates microfilm 5017, 62-214 IIIF image files per item (sampled 4) at tile.loc.gov. HMML: EMML microfilm of ~9,000 Ethiopian manuscripts (HMML's own figure; ~7,500 reels shipped to HMML) inside the ~75,000-manuscript Eastern Christian collection - all images login-gated. Sinai Palimpsests: 160+ palimpsests with spectral imaging, research site login-gated; public galleries contain no Ethiopic material. Internet Archive holds essentially no Ge'ez manuscript scans (2 hits for language:Geez, both objects/amulets; 290 title hits for ethiopic/geez are European scholarship about the language).

Licence: not stated

Access: No image bulk download or dump. Closest: (1) GitHub tree API for the whole TEI corpus (210 MB repo; git clone works), (2) one manifest per record then per-canvas image fetches, (3) for the other IIIF hosts, one manifest per shelfmark from their IIIF collection endpoints (CUDL /iiif/collection/ethiopicmanuscripts, Bodleian /iiif/collection/ethiopian-and-eritrean). DigiVatLib enumerates via https://digi.vatlib.it/mss/ajax/category/<fond> (JSON). HMML offers a Data Portal dataset export (/dataPortal/export/full, /dataPortal/export/listing) but both need a signed-in account. · sample HTTP 200

Not reachable: https://www.vhmml.org/readingRoom/search - 403 'Turnstile validation failed' (Cloudflare Turnstile); no anonymous search or image access, requires account; https://www.vhmml.org/dataPortal/search - 500 'Unable to search the Data Portal'; /dataPortal/export/listing 400 without a session; https://sinai.library.ucla.edu/browse - 200 but body 'Unauthorized. Please login to access this page' (registration required); https://sinaipalimpsests.org/ - TLS certificate is for anubispoc.library.ucla.edu (InCommon/UCLA), so HTTPS fails name verification; plain http:// returns 200 but the site is a static archive with no viewer or IIIF; https://www.loc.gov/item/<id>/manifest.json - 403 (item ?fo=json works and exposes the tile.loc.gov image URLs); https://digital.soas.ac.uk/ - 403 Forbidden to this host (UA and Accept-Language changes did not help); https://emip.arts.kuleuven.be/ - DNS does not resolve; emip.mml.ox.ac.uk, emip-ms.org, ethiopianmanuscripts.org all NXDOMAIN (the EMIP project site proper is gone; its scans survive as the Beta Masaheft IIIF above and in vHMML); Gallica/BnF - 403 from this host (known from earlier exploration); BnF Ethiopic mss reachable only via archivesetmanuscrits.bnf.fr links, not tested to completion here; https://betamasaheft.eu/iiif/<id>/manifest and /manifest.json - 404 '/media/aai-hlz-bm is neither a file nor part of an image sequence'; the working manifest path is /api/iiif/<id>/manifest

Working path, in order. (0) Always send a browser User-Agent and the cookie 'ds_verification_token=proof_passed' to betamasaheft.eu: without it the server returns 200 with a <script> that sets the cookie and reloads, i.e. an empty body that looks like a successful fetch (that is why plain curl look…

Gallica / BnF

General catalogue (catalogue.bnf.fr SRU, live counts 2026-11-13): 1,679,003 records flagged bib.digitized=freeAccess (Gallica, free online). By item language: Coptic (cop) 459 records of which 12 digitised (1 manuscript); Ge'ez (gez) 329 records of which 27 digitised (3 manuscripts); Amharic (amh) 236 records of which 11 digitised (1 manuscript); Greek (grc) 143,420 records of which 126,039 digitised freeAccess (+4,503 restrictedAccess, Gallica intra muros). Digitised manuscripts overall (doctype=t) 37,073, of which Greek 770 of 777 Greek manuscripts in the catalogue. Coptic digitised set (12, Gallica ids): bpt6k91092706 = A compendious grammar of the Egyptian language (Tattam, 1830); bd6t5253301q = كتاب دورة عيدي الصليب والشعانين (Coptic-Arabic liturgy, Arabic script); bpt6k904189d = [Bible copte], 1811; bpt6k91169840 = Fragment des révélations apocryphes de saint Barthélemy, 1835; bpt6k54538305 = Histoire du patriarche copte Isaac, 1890; bpt6k5696685z = Oeuvres de Schenoudi I (1907-1914); bpt6k5604078n = Oeuvres de Schenoudi II (1907-1914); bpt6k9063d = Les Évangiles des douze Apôtres et de saint Barthélemy, 1946; bd6t53824501 = Grammaire copte, 1907; bpt6k318523d = Alphabetum cophtum sive aegyptiacum; bpt6k108650z = Chronique de Jean, évêque de Nikiou, 1883; bpt6k75521z = Vie et récits de l'abbé Daniel le Scétiote (VIe s.). Ge'ez manuscripts digitised: btv1b102246940 (texte biblique en guèze), btv1b10891625b (code éthiopien, 340 p., early 18th c.), btv1b10892850f (ወንጌለ ዮሐንስ, Gospel of John); Ge'ez printed: btv1b55005346w (Psalterium David, 1513), bpt6k14301981 (Rituel éthiopien), bpt6k33073h (Chronique de Galâwdêwos, 1895), bpt6k3153038 (Livre des mystères, 1903), synaxaire éthiopien volumes bpt6k9085k/bpt6k9124v/bpt6k90965. Europeana (provider='National Library of France', live): 377 items match 'copte', of which 104 titles are the manuscript series 'Copte N' with 100 distinct numbers (Copte 10-103+); 875 items match 'éthiopien', of which 207 distinct 'Éthiopien N' manuscripts (numbers 34-756); 58 items match 'guèze'. Caveats: the manuscript department's own inventory (Archives et manuscrits, archivesetmanuscrits.bnf.fr) is 403 from here, so catalogue SRU undercounts manuscript series (e.g. only 1 Coptic manuscript record in the general catalogue vs 100+ Copte N items visible on Europeana). Many 'Coptic' catalogue hits are Western philological editions in French/German/Latin, not Coptic text.

Licence: not stated

Access: none reachable. Documented Gallica SRU caps maximumRecords at 50; IIIF would be page-by-page; no bulk archive/API responds. · sample HTTP 403

Not reachable: https://gallica.bnf.fr/* - 403 Cloudflare 'Attention Required' (WAF), from this VPS AND from gx10, over IPv4 and IPv6, with curl, Chrome UA and headless Chromium; /cdn-cgi/trace returns 200 so the edge is reachable and the block is a firewall rule.; https://gallica.bnf.fr/SRU, /iiif/.../manifest.json, /ark:/.../texteBrut, /ark:/.../f1.item - all 403, same rule.; https://archivesetmanuscrits.bnf.fr/ - 403 from Apache (BnF Archives et manuscrits, the manuscript-department inventory).; https://iiif.bnf.fr/ and https://api.gallica.bnf.fr/ - DNS does not resolve.; https://openapi.bnf.fr/ - WSO2 APIM gateway answers 200 but exposes no Gallica API (all /api/am/* paths 404).; https://web.archive.org/cdx for gallica texteBrut / IIIF paths - no snapshots.

Discovery works today and is enough to build the dossier registry: (1) query catalogue.bnf.fr SRU with bib.digitized all "freeAccess" combined with bib.language all "cop|gez|amh|grc" and optionally bib.doctype all "t" for manuscripts, maximumRecords 50, startRecord paging - no key, no rate limit ob…

greek-ms: Greek manuscript repositories (Athos, Sinai, Patmos, Jerusalem, Laurentian, Ambrosian, Paris, Munich) - open IIIF and catalogue APIs for patristic manuscripts

What is actually reachable and scanned, measured today. (a) Mount Sinai, St. Catherine's: LOC collection 'manuscripts-in-st-catherines-monastery-mount-sinai' = 1,691 digitised items (facets: 'Original Format: Manuscript/Mixed Material 1,685, Web Page 6'; 'Online Format: Image 1,690'), broken down by the microfilm series in the 'Part of' facet: Greek 1,078, Arabic 312, Syriac 156, Georgian 86, Slavonic 41, Ethiopic 6, Turkish scrolls 4; language facet 'Greek, Ancient (to 1453) 1,081', Arabic 308, Syriac 156, Georgian 85, Church Slavic 40. These are the 1949-50 LOC/AMERICAN Foundation expedition microfilms (film numbers 5010-5016), digitised from film, so page images are monochrome. (b) Jerusalem, Greek and Armenian Patriarchates: LOC collection = 1,009 items; Part-of facet: Hagios Sabas 272, Panagios Taphos 253, Georgian 126, Arabic 105, Hagios Stavros 90, Armenian 31, Syriac 26; language facet: Greek 399 + Greek (modern) 270 = 669, Georgian 125, Arabic 96, Armenian 32, Syriac 27, Geez 20, Slavic 20. Also film-derived (microfilm 5017-5018). (c) Sinai palimpsests: the UCLA/EMEL electronic library (sinai.library.ucla.edu) holds spectral images of 'more than 160 palimpsests' (project page) but the browse interface is behind registration - no anonymous count possible. (d) Biblioteca Medicea Laurenziana: the real digital library is NOT teca.bmlonline.it (unreachable) but a CONTENTdm site, tecabml.contentdm.oclc.org, with 13 collections; collection 'plutei' = 4,152 objects (the Plutei, which is where the Greek mss Plut. 1-87 live), 'orientali' and 'med-palatino' separate. (e) Bayerische Staatsbibliothek: IIIF works per identifier but the catalogue search is broken from this host, so no Greek-manuscript count could be measured; the IIIF top collection reports total 3,222,873 manifests (prints + mss, unfilterable). (f) Biblioteca Apostolica Vaticana (not on the brief's list but the largest Greek holding): DigiVatLib IIIF works, e.g. Vat.gr.1205 manifest = 150 canvases; the library states its programme aims to digitise '80,000 codices'. (g) Ambrosiana, BnF, Athos, Patmos: no scanned object could be fetched - see blocked[]. Catalogue-only coverage is strong and is the real discovery layer: Pinakes/IRHT holds at least 80,000 manuscript notices (id 80,000 resolves to a real notice; 82,500 returns the home page instead, so the ceiling is between the two - the ids are Diktyon numbers and are not strictly contiguous), one per Diktyon number, and every notice carries the external links (NT.VMR, DBBE, Princeton, Marciana, Atticisme) but frequently NO digitisation link - e.g. Hagion Oros, Moni Dionysiou, Mss. 0032 (id 20000) has an empty 'Visualiseur' field, i.e. no online facsimile.

Licence: not stated

Access: Per repository, no bulk archive. LOC: enumerate the two collections with c=1000-style paging and read resources[].files[].url, or pull the JP2 directly from https://tile.loc.gov/storage-services/service/amed/amedmonastery/<item-id>/<NNNN>.jp2 (verified 200 image/jp2, 1.32 MB for 00271073823-jo/0001.jp2). BSB: GET https://api.digitale-sammlungen.de/iiif/image/v2/<bsb-id>_<page>/full/full/0/default.jpg (verified 200). Laurentian: /digital/api/collection/plutei/id/<pointer>/download for the whole object, or per-page IIIF (verified 200, 243 KB at 800px, 2.5 MB at full). Vatican: https://digi.vatlib.it/iiifimage/<MSS_shelfmark>/<file>.jp2/<region>/<size>/0/default.jpg (verified 200). · sample HTTP 200

Not reachable: https://teca.bmlonline.it/ - connection timed out twice (25 s and 55 s, 0 bytes). The Laurentian digital library has moved to CONTENTdm (https://tecabml.contentdm.oclc.org), reachable and usable; the old host is dead from here.; https://gallica.bnf.fr/ - HTTP 403 on every path tried (IIIF manifest for ark:/12148/btv1b52505629f, the ark HTML page, the SRU API, the accueil page). Site-wide block of this host, not a per-resource restriction.; https://archivesetmanuscrits.bnf.fr/ - root returned 200 once, then 403 on every subsequent request (record path ark:/12148/cc1000p and /recherche.html). BnF is effectively unreachable from this machine; Paris coverage could not be measured.; https://ambrosiana.comperio.it/ - every path returns HTTP 200 but the body is a 16,913-byte JS challenge ('Verifica in corso... / Checking...'); /iiif/2/<id>/info.json and /opac/iiif/2/<id>/manifest return the same challenge page, so the Ambrosian IIIF endpoint could not be verified. https://iiif.ambrosiana.it does not exist (TLS name mismatch).; https://www.mountathos.org/ - HTTP 403 on both http and https with a browser User-Agent (5.4-5.6 KB block page). No Mount Athos digital library reachable. athoslibrary.gr is an Athonite bibliography blog whose PDFs are Google Drive links, not a manuscript repository; athosfriends.org (200) carries no digitisation or IIIF content.; https://www.patmosmonastery.gr/ - HTTP 200 but the site is a parish/monastery site: it has a library page and a manuscript conservation workshop page, no digital library, no IIIF, no catalogue API. www.patmiaki.gr and www.patmoslibrary.gr do not resolve.; https://www.digitale-sammlungen.de/ (Munich) - all search and browse URLs return HTTP 200 with the body 'An error occurred while serving your page' (/en/search?query=Homer, /api/search, /en/browse/handschriften -> 404); the OAI endpoints (api./oai, oai.digitale-sammlungen.de, services./sru, sru./sru) are 404 or NXDOMAIN. The IIIF APIs at api.digitale-sammlungen.de work normally, so navigation is only possible by identifier.; https://pinakes.irht.cnrs.fr/api/depots/liste_villes/ and /api/depots/liste_fonds_by_depot/<id>/ and /api/cotes/liste_cotes_by_fond/<id>/ - all return a 500 'IRHT - Service temporairement indisponible' sentinel-guard page, as does POST /recherche-generale/results/page. Individual notice pages (/notices/cote/<id>/) are served normally, so Pinakes is browse-only from here.; https://digi.vatlib.it/search?q=Vat.gr - HTTP 403 (bot block); /view/<shelfmark> and the IIIF endpoints are 200. No count of digitised Vat.gr. mss could be obtained.; https://www.internetculturale.it/ - search returns a 'Firewall Captcha Authentication' page; the Italian aggregator (which would have covered Laurentian/Ambrosiana) is unusable.; https://diktyon.net/ - no longer the Greek manuscript database: the domain now serves an unrelated personal site (Astro build, LinkedIn links). Any reference to 'diktyon.net API' is stale; Diktyon numbers survive only as the id field inside Pinakes.; https://sinaipalimpsests.org/ - TLS certificate does not match the hostname (curl 60) on both apex and www; http:// returns 200 and redirects the reader to sinai.library.ucla.edu, which requires registration ('Unauthorized. Please login to access this page.').

1) Decide the layer first. Printed Greek patristic text -> use the IA Migne route (sources-scans-ia-migne.json), not this source, because only 29 of 165 PG tomes have usable Greek OCR and 137 have none. Manuscript images -> this source, with HTR, not OCR. 2) LOC is the highest-value, lowest-frictio…

HathiTrust + Google Books + e-rara + MDZ (Bayerische Staatsbibliothek) + ÖNB (Austrian National Library) — public-domain full-view scans

HathiTrust: the Bibliographic API resolves identifiers to htid + item URL (verified: oclc/424023 -> record 000578050, htid mdp.39015025315527; oclc/6344592; oclc/1085860). Corpus size not measurable from this host (hathifiles 403); HT holds the standard Migne PG/PL runs and most pre-1929 Greek, Coptic and Ethiopic editions, but none of that is reachable here (see blocked). Google Books: 380,560 Google-digitised volumes are mirrored on Internet Archive as bub_gb_* (IA advancedsearch identifier:bub_gb_*); of those 1,515 are language grc/gre/el, 7 are Coptic (language:cop), 14 are Amharic/Ge'ez/Tigrinya/Syriac (language:amh OR gez OR tir OR syr) — this mirror, not Google itself, is the usable Greek/Coptic/Ethiopic surface. e-rara (Swiss rare books, all pre-1900 imprints): 119 records match 'griechisch' in the title, of which 89 are flagged OCR-searchable; 10 match 'Graecus'/'Septuaginta'; 0 for 'Koptisch'/'Aethiopisch'/'äthiopisch'/'Migne'/'Patrologia' — e-rara has no Coptic or Ethiopic material and no patrology runs. MDZ: 422-page bilingual Greek-Latin Jesuit volume bsb12188301 and 8,436-page e-rara-style runs are indexed; Greek manuscripts and early Greek prints are present (verified bsb12188301, bsb12188527), but the site exposes no query endpoint (JS SPA, /api/search -> 400, oai-pmh -> 404), so corpus-wide Greek counts must come from the BSB catalogue. ÖNB: IIIF v3 + REST work on any 8-hex object id (verified 1008BEB0, 106F8E6A = Vienna Genesis, 53 folios, Greek); discovery is broken from here (onb.digital -> 504), so coverage cannot be counted.

Licence: not stated

Access: None usable from this host: HathiFiles (https://www.hathitrust.org/hathifiles) -> 403; MDZ OAI-PMH retired (oai-pmh -> 404, sru.digitale-sammlungen.de does not resolve); e-rara OAI ListRecords works but only 10 records per page with resumption tokens; ÖNB has no bulk endpoint. Largest practical bulk surface is the Internet Archive mirror of Google scans (identifier:bub_gb_* -> 380,560 items via IA advancedsearch). · sample HTTP 200

Not reachable: babel.hathitrust.org (all /cgi/pt, /cgi/htd, /cgi/imgsrv paths) -> HTTP 403, Cloudflare managed challenge ('cf-mitigated: challenge', 'Just a moment...'), reproducible with browser UA; no JS challenge solver on this host.; catalog.hathitrust.org/Record/<id> -> 403 (same Cloudflare challenge); only /api/volumes/* answers 200.; www.hathitrust.org/hathifiles -> 403; no bulk metadata or Extracted Features without HTRC account.; www.googleapis.com/books/v1/volumes -> 429 RESOURCE_EXHAUSTED, daily quota 0 for the anonymous project (624717413613).; books.google.com/books?...&output=text -> 403 'Sorry...' (bot interstitial).; books.google.com/books/content?id=…&img=1&zoom=3 -> 200 but returns a fixed 9,103-byte placeholder PNG on preview-only volumes.; onb.digital (search + result pages) -> 504 Gateway Time-out; onb.digital/api/search -> 404.; pdf-service.onb.ac.at/<ID> -> connection reset (000) on every attempt.; api.onb.ac.at /rest/v1/ie/* -> intermittent 'Recv failure: Connection reset by peer' (000), ~1 in 3 requests.; MDZ discovery: /api/search -> 400/406, /oai-pmh -> 404, sru.digitale-sammlungen.de and bdr.bib-bvb.de do not resolve; onb.alma.exlibrisgroup.com, onb.primo.exlibrisgroup.com and sru.bib-bvb.de do not resolve from this host — no way to enumerate Greek/Coptic/Ethiopic holdings programmatically.; e-rara: /i3f/v20/<DOI-number>/manifest and /download/pdf/<VLID> (no domain prefix) -> 404; identifiers must be the internal VLID with its domain prefix.; download.digitale-sammlungen.de/BOOK/<bsb>/<bsb>.{pdf,txt} -> 404 (legacy download host retired).

1) HathiTrust: unusable for text. The Bibliographic API is only an identifier resolver: GET https://catalog.hathitrust.org/api/volumes/brief/oclc/<oclc>.json -> records[].htid + items[].itemURL; use it to map OCLC/ISBN/LCCN to htid, then look the same edition up on Internet Archive (many htid scans…

Internet Archive — Corpus Scriptorum Christianorum Orientalium (CSCO) and oriental Christian editions

One multi-file item (Corpus-Scriptorum-Christianorum-Orientalium) holds 115 scanned CSCO volumes = CSCO main series 1-155 with 15 gaps (124, 125, 128, 129, 131, 135-142, 145, 146). Those 115 bindings carry 140 main-series numbers and 138 sub-series numbers: Scriptores Syri 71, Aethiopici 30 (1-30, i.e. the first 30 of 117), Coptici 16 (1-14, 19, 20), Arabici 14 (1-14), Iberici 4, Subsidia 3 (2, 4, 8). Separate items add ~10 more volumes — CSCO 367/375/376 (Synodicon in the West Syrian Tradition I-II, 1975-76), 439/440 (Canons of Maruta, Vööbus 1982), 612/613, 650, 656/657 (Syr 253) — for ~150 distinct volumes of the 713 published (~21%). Against the published sub-series that is Coptici 16/55 (29%), Aethiopici 30/117 (26%), Arabici 14/53 (26%), Syri 71/265 (27%). Subject matter is strong where the project needs it: the Coptic block is exactly the Shenoute/Pachomius/Anthony core (Sinuthii vita et opera = CSCO 41, 42, 73, 96, 108; S. Pachomii vitae = 89, 99-100, 107; S. Antonii vita = 117, 118; Athanase lettres festales = 150, 151; Acta Martyrum = 43, 44, 86); the Ethiopic block is the Conti Rossini/Guidi hagiography-and-chronicle corpus (Vitae Sanctorum, Acta Martyrum, Liber Axumae, Annales Iohannis/Iyasu/Bakaffa, Walatta Petros); the Arabic block is the editio princeps run for Eutychius, Agapius, Severus ibn al-Muqaffa', Petrus ibn Rahib and the Synaxarium Alexandrinum. Two further items are dark: csco-368 and csco-432_syr_188… resolve at /metadata with an empty file list and return HTTP 404 at /details.

Licence: not stated

Access: Per-file only, no bulk API and no archive torrent (item metadata sets noarchivetorrent:true). Files: https://archive.org/download/Corpus-Scriptorum-Christianorum-Orientalium/<url-encoded name> → 200. Total item 20.29 GB, of which 115 image-container PDFs = 1.86 GB and 115 per-volume JP2 zips ≈ 18 GB; 115 _djvu.txt (56 MB total) and 115 _djvu.xml (e.g. 3.5 MB for CSCO 152) are the cheap text/geometry paths. No API key, no login, no rate limit hit at 4 concurrent requests. · sample HTTP 200

Not reachable: https://ia-fts.archive.org/api/v1/search/hits?… — connection failed (curl exit with no response, code 000): IA's search-inside API is not reachable from this host, so full-text search over the un-OCR'd volumes is unavailable.; https://iiif.archive.org/iiif/<id>$<leaf> — HTTP 500 for leaf >= 200 on the 115-volume multi-file item (200 only for leaves 0-100).; https://archive.org/details/csco-368 and https://archive.org/details/csco-432_syr_188_theodore_bar_koni_livre_de_scolies — HTTP 404 (dark); /metadata still answers with an empty file list.; https://archive.org/download/<id>/page/nN — returns the same leaf-0 image for every N on this multi-file item, so per-page fetch is unusable.

1) Manifest: curl https://archive.org/metadata/Corpus-Scriptorum-Christianorum-Orientalium > meta.json, then select files with format Text PDF / Image Container PDF (115, 1.86 GB total) plus the parallel DjVuTXT and Djvu XML entries. 2) Download by exact filename with Python urllib.parse.quote — na…

Internet Archive (archive.org) — scan discovery mechanics

Everything is a page-image scan with a derived text layer; nothing is born-digital. Migne Patrologia Graeca: collection pub_patrologia-graeca = 162 items (= the 162-volume series; GTU microfilm, language mislabelled 'English'), collection pub_patrologia-latina = 218 items (Latin series). Separately 281 items match title:"patrologiae cursus completus" AND language:grc (Google/unknown-library scans of the same volumes, many duplicates). Greek-language holdings overall: language:grc 27,062 items; language:gre 154,931. Coptic: language:cop 410 items (mostly modern editions/translations, not primary Coptic text). Ethiopic: language:eth 56 items; Amharic language:amh 1,312; Tigrinya language:tir 155; Syriac language:syr 4,632. Sampling: 120 Coptic items -> 103 had _djvu.txt; 56 Ethiopic items -> 41 had _djvu.txt. NOTE: the exploration note 'collection:patrologia-volumes has 220 PDFs' does not hold in the public index — q=collection:patrologia-volumes returns numFound 0 (HTTP 200); the 218-item pub_patrologia-latina looks like the same body of scans under a different collection name.

Licence: not stated

Access: No bulk API for page images; per-file GETs only. Text can be pulled whole (_djvu.txt, 9.3 MB for PG 56) or by HTTP Range (206, verified). Metadata for many items: scrape API. JP2 page images are single-member HTTP GETs out of _jp2.zip (~1.5 MB/page); _jp2.tar items (e.g. PatrologiaGraecavolume56, 2.35 GB) have no per-page path other than page/nX.jpg or IIIF. · sample HTTP 206

Not reachable: https://ia-fts.archive.org/api/v1/search — DNS failure (getent hosts: no such host), HTTP 000; the working full-text endpoint is the services/search/beta/page_production one; https://archive.org/metadata/<id>/bookreader — 200 but {"error": "Couldn't get 'bookreader' for item ..."}; https://iiif.archive.org/iiif/3/<id>/page/n<N>/full/1000,/0/default.jpg and /iiif/<id>/page/... — HTTP 404 (207-byte HTML error); only the URL-encoded image/iiif/3 form works; https://<server>/BookReader/BookReaderImages.php?zip=...&file=..._jp2/..._0120.jp2 — HTTP 422 body 'Invalid zip' on a _jp2.zip item, 404 on the other path form; https://archive.org/download/PatrologiaGraecavolume56/PatrologiaGraecavolume56_jp2.zip/... — HTTP 404, item only has _jp2.tar (2.35 GB); https://archive.org/download/apocryphonjohann0000unse/apocryphonjohann0000unse_djvu.txt — HTTP 401 '401 Authorization Required' (access-restricted-item: true, printdisabled/inlibrary); https://archive.org/details/PatrologiaGraecavolume56 — renders no rights/copyright statement at all (grep for 'public domain', 'no known copyright', 'in copyright': 0 matches); rights are simply not asserted for that item

1) Discover: advancedsearch.php with fl[]=identifier,title,language,collection and output=json; use scrape API for larger sets (count>=100). 2) Detect OCR before promising text: GET /metadata/<id> and look for *_djvu.txt, *_djvu.xml, *_hocr.html, *_chocr.html.gz, *_jp2.zip vs *_jp2.tar, and access-…

ia-migne: Migne Patrologia Graeca and Patrologia Latina scans on the Internet Archive

Two corpora on IA, both near-complete, but they are many separate items rather than one collection. PATROLOGIA LATINA (PL, 221 vols): 944 items match the identifier families patrologiaecursu*/patrologiaecur*/patrologiaecurs* (~1,128 GB of item storage). 242 distinct volume numbers are represented, of which 220 of the 221 volumes of PL 1-221 are present — only vol. 185 has no scan — plus 22 numbers above 221 (222-275, supplements and copies numbered outside the run). Duplication is heavy: 47 volume numbers have 2+ items (e.g. vol. 4 has 6: patrologiaecursu0004mign, ..._g2j9, ..._r7r0, ..._s6m4, patrologiaecursu04mignuoft, patrologiaecursu04hammuoft). Scans come from three digitizers — Claremont School of Theology (177 items), University of Toronto/Robarts (76), PIMS Library (59). Separate PL containers: item PatrologiaLatina holds 85 Text PDFs for vols 1-50, 52-81, 90-93, 95 (99.6 GB, 85 PDFs + 51 JP2 zips, and ZERO DjVuTXT text derivatives); item patrologia-latina_1-221 is titled 'Patrologia Latina vol. 1-221' but contains exactly one 22 MB PDF named 'Patrologia_100,103.pdf' — i.e. vols 100 and 103 combined — a 3-title-bait; do not treat it as a 221-volume set. PATROLOGIA GRAECA (PG, 161 vols in 166 tomes): the container item patrologia-volumes ('Patrologia Graeca PDF Volumes', uploader evagelosvar@gmail.com, 2019-07-27) holds 165 tomes numbered 001-161 with the a/b/c splits (7a,7b,16a,16b,16c,86a,86b,87a,87b,87c): 155 distinct tome numbers = PG 1-161 minus vol. 75 and vol. 106, which are simply absent from this item. 2,042 files / 347.8 GB: 165 DjVuTXT, 165 DjVu XML, 165 Single-Page JP2 ZIPs (2.1 GB for vol. 56), 165 hOCR + chOCR + EPUB + Daisy + OCR Page Index + page_numbers.json, 110 Text PDFs, 55 Image Container PDFs. Independent PG item families: patrologiae_cursus_completus_gr_vol_* = 78 items / 122 GB (300 ppi, uploader andrewmaguire57, 2015, tesseract '-l grc'), patrologicursusNNNmigngoog + patrologiaecurs*goog Graeca = part of 665 Google-scanned items / 478 GB (600 ppi, '-l lat+grc' or '-l ell+lat'), plus PatrologiaGraeca100, PatrologiaGraeca.Indices, 'Patrologia Graeca (volume 56)', 3 items Patrologia_Graeca_vol_*, and patrologiae_cursus_completus_patrologiae_graecae_tomus_clvii_1866. Across those alternate families 116 distinct PG volume numbers appear, including the two (75, 106) missing from the container item — so PG 1-161 is fully covered across families, but not by any single item. NOT on IA: no PG 162 (the general index) except PatrologiaGraeca.Indices, and no Coptic/Ethiopic/Arabic Migne material beyond the Graeca/Latina runs.

Licence: Public Domain Mark 1.0 where the item states a licence at all (quoted from the patrologiae_cursus_completus_gr_vol_042 page: rel="license" title="Public Domain Mark 1.0" href="htt…

Access: Per volume, not per corpus: https://archive.org/download/<id>/<id>_djvu.txt returns 206 to a Range request (verified on 6/6 sampled Latina items and on the Graeca items). Full file list + md5s: <id>_files.xml. No single archive of the whole corpus exists; the only bulk-ish artefact is the container item patrologia-volumes, whose 2042 files total 347.8 GB and are fetched one file at a time. The `ia` CLI (pip install internetarchive) is the convenient driver: `ia download <id> --glob='*_djvu.txt'`. · sample HTTP 206

Not reachable: collection:patrologia-volumes — q=collection:"patrologia-volumes" on advancedsearch.php returns 200 with numFound 0: no such collection exists, patrologia-volumes is a single item; the '220 PDFs' figure in earlier notes is wrong (2042 files, 165 books, 110 Text PDF + 55 Image Container PDF).; https://iiif.archive.org/iiif/patrologia-volumes/info.json — HTTP 404 (wrong IIIF shape); the working endpoints are /iiif/3/<id>/manifest.json (200) and /image/iiif/3/<id>%2Fpage%2Fn<N>/full/max/0/default.jpg (200).; patrologiaecurs104migngoog — <id>_jp2.zip returns HTTP 404 (Google-scanned items often have no JP2 zip), while <id>.pdf and <id>_djvu.txt return 206; the image path for those must come from the PDF or the IIIF manifest's own canvas image ids.; patrologia-latina_1-221 — not blocked but mislabelled: title 'vol. 1-221', contents 'Patrologia_100,103.pdf' only.; Gallica (BnF) — HTTP 403 from this host (carried over from prior reconnaissance, not re-tested here); use IA copies instead.

1) Map, do not crawl. Build the volume index from advancedsearch with fl[]=identifier,volume,item_size over the three PL id families, then reconcile to a canonical vol. 1-221 number; expect 2-6 duplicate items per volume and pick ONE per volume (prefer the 2025/Claremont tesseract 5.3 derivatives o…

OCR toolchain for Greek, Coptic, Ethiopic and Arabic scripts (Tesseract grc/lat/ell/cop/ara/amh/tir/syr, Kraken, OCR4All/Calamari, Coptic Scriptorium and KELLIA models, IIIF page delivery)

Toolchain coverage measured on live scans, not claimed: Tesseract 5.3.4 with 10 packs installed (amh, ara, chi_sim, ell, eng, grc, lat, osd, syr, tir) plus a community Coptic pack (cop, 7.7 MB float model). Scan corpus used for testing: IA item patrologia-volumes = 165 Migne Patrologia Graeca volumes (155 numbered 001-161 with gaps 007/016/075/086/087/106, plus lettered 007a/b, 016a/b/c, 086a/b, 087a/b/c), 220 PDFs, 165 DjVuTXT + 165 DjVu XML; separately uploaded PatrologiaGraecavolume56 (grc OCR, 9.34 MB djvu.txt) and PatrologiaGraecaVolume65 (grc+lat OCR, 4.66 MB djvu.txt) with full-resolution page JPEGs (4102x6322 and 4077x6199). IA at large: 26,505 texts tagged language:grc, 376 cop, 178 gez (advancedsearch numFound 2026-11-13). Project-facing need: 344 gap dossiers, 102 with no e-text (59 Greek, 36 Coptic, 7 Ethiopic; 242 have an e-text, 93 of those carry an archive.org link = 102 links). NB the task brief said 330 dossiers/101 without e-text; the live gaps.json (meta.json generated 2026-11-13T22:35:13Z) says 344/102 — use the live numbers. OCR reality in the bulk item: 18 volumes sampled (7 full djvu.txt, 11 mid-file 600 KB slices) and only 4 (050, 060, 150, 160) contain any Greek OCR; the rest are 0% Greek even where the pages are Greek (volume 001 pages n150/n350/n450 OCR to Greek script under grc at lexicon 0.525-0.562, readable but error-heavy, while its djvu.txt has 0 Greek characters because IA ran -l que+lat+Latin).

Licence: not stated

Access: https://archive.org/download/patrologia-volumes/{001..161}{a,b,c}_djvu.txt and _djvu.xml (165 text derivatives, 155 numbered + 10 lettered volumes, 220 PDFs); page derivatives via https://archive.org/download/patrologia-volumes/{vol}_jp2.zip/{vol}_jp2%2F{vol}_NNNN.jp2 (URL-encode the slash: unencoded path returns 503) · sample HTTP 200

Not reachable: https://gallica.bnf.fr/ark:/12148/btv1b10722462x -> 403 from this host (BnF Gallica blocks the request; no IIIF sample obtained); pip install pylaia -> ResolutionImpossible (dependency conflict with the installed torch 2.11.0 / numpy 1.26.4 stack); PyLaia not tested; https://zenodo.org/api/records?q=coptic+ocr -> 403 (Zenodo API refused; community model registry could not be queried); boaztulu/hhd-ethiopic-ocr: README claims CER 12.68/21.88, but checkpoints/crnn_vgg_s2 contains only config_used.yaml and metrics.jsonl; no weights and no GitHub releases, so the model is not downloadable/testable; https://archive.org/download/patrologia-volumes/001_jp2.zip/001_jp2/001_0150.jp2 (unencoded member path) -> 503; the URL-encoded form .../001_jp2%2F001_0150.jp2 -> 200 (pitfall, not a hard block); https://ia800703.us.archive.org/BookReader/BookReaderImages.php?zip=...&file=... (without id=) -> 404; adding id=patrologia-volumes -> 200 (pitfall); GitHub API returned 403 rate-limit on repository search; raw.githubusercontent.com and per-repo endpoints still answered

1) Resolve the edition to an IA identifier (advancedsearch title:"Patrologia Graeca"; 12 items) and read metadata before downloading anything: ocr_parameters is the tell. Seen in the wild: -l que+lat+Latin (patrologia-volumes, 165 volumes, detected Fraktur/Latin/Arabic), -l grc (PatrologiaGraecavol…

Wikisource Ge'ez texts and public-domain translations

LIVE COUNT 2026-11-13, from page lists and byte/word counts of the fetched wikitext, not from descriptions. There is NO Ge'ez-language Wikisource: gez.wikisource.org does not resolve, 'gez' appears in none of the 84 wikisource subdomains in the Wikimedia sitematrix (smstate=all, so never existed in any state), gez.wikipedia.org and gez.wiktionary.org likewise do not resolve, and dumps.wikimedia.org/gezwikisource/ returns 404. The multilingual wikisource.org holds nothing Ge'ez: action=query search returns 0 hits for 'Ge'ez' and 0 for the Ge'ez-script string ውሕዝ, and 3 hits for 'Ethiopic', all Interlingua Bible pages. What actually exists on en.wikisource is a small set of ENGLISH TRANSLATIONS of Ge'ez-base texts: (a) The Book of the Saints of the Ethiopian church (Budge 1928, the Ethiopic Synaxarium) - 13 month pages, 12 present (Tir, Yekatit, Megabit, Miyazia, Ginbot, Senne, Hamle, Nehasse, Pagumen, Meskerem, Tekemt, Hedar = 454,707 words / 2.50 MB wikitext measured); the 13th month, Tahsas, has no page, so ~1/13 of the cycle is missing. (b) The Book of Enoch (Charles 1917) - 110 subpages, 39,876 words / 236,479 chars, essentially complete. (c) Kebra Nagast (Budge 1922) - title page + 1,107 words of text, flagged {{incomplete}}; this is roughly 1% of the ~90,000-word translation (estimate from the print edition), i.e. a stub, not a text. (d) First Book of Ethiopian Maccabees / 1 Meqabyan (D. P. Curtin 2018) - 885 words, partial. (e) Translation:1 Meqabyan (Wikisource user translation) - 7,689 words. Total measurable English text across all five: ~504,000 words. GE'EZ-SCRIPT PRIMARY TEXTS: 0 - a search for the Ge'ez word መጽሀፈ returns exactly one page, the header of 1 Enoch, which merely cites the Ge'ez title. Only 2 pages in the whole wiki declare Ge'ez as source language (insource:"language = gez" -> 1 Enoch, 1 Meqabyan); the Maccabees page in fact declares language = am. Category:Ethiopian Orthodox Church has 8 members, of which 2 are Ge'ez-derived texts. No Ethiopic-script corpus, no critical editions, no apparatus.

Licence: Two layers, both verified on live pages. (1) Underlying works: public domain. The Synaxarium parent page carries {{PD-US}} and {{PD/US|1934}}; Module:PD-US reads args[1] as the AU…

Access: POST https://en.wikisource.org/wiki/Special:Export with form field pages=<titles separated by %0A>, curonly=1, action=submit -> one request returned 18 pages / 2.55 MB (verified 200). Enumerate subpages with /w/api.php?action=query&list=allpages&apprefix=<Root>/&aplimit=500. Whole-wiki fallback: https://dumps.wikimedia.org/enwikisource/latest/enwikisource-latest-pages-articles.xml.bz2 (HTTP 200, Content-Length 3,438,896,364 = 3.44 GB, Last-Modified 2026-11-01) - far too big for 0.5 MB of relevant text; use Special:Export instead. · sample HTTP 200

Not reachable: https://gez.wikisource.org - DNS does not resolve (curl exit 6, 'Could not resolve host'). Not a firewall on this machine: wikisource.org and am.wikisource.org both resolve from here, so the host simply does not exist.; gez.wikipedia.org and gez.wiktionary.org - same, no DNS. 'gez' has no Wikimedia project in any state.; https://dumps.wikimedia.org/gezwikisource/ - HTTP 404 (dir absent); https://dumps.wikimedia.org/gezwiki/ - HTTP 404.; incubator.wikimedia.org Ws/gez - list=allpages&apprefix=Ws/gez returns 0 pages (no Ge'ez Wikisource draft in the Incubator either).; wikisource.org (multilingual) - 0 hits for 'Ge'ez', 0 for ውሕዝ, 3 hits for 'Ethiopic' and all three are Interlingua Bible text, i.e. no Ge'ez corpus.; am.wikisource.org - redirects to the multilingual wikisource.org (there is no Amharic Wikisource), and the Ge'ez article title tried there gives 404.; Transient: en.wikisource.org API returned HTTP 429 on bursts (retry-after 2, x-envoy-ratelimited); resolved by pacing requests >=3 s with a descriptive User-Agent. Not a block on this machine.

1) Resolve the tree first: for each root page call list=allpages&apprefix=<root>/&aplimit=500. Today that yields Kebra Nagast 0 subpages, 1 Meqabyan 0, First Book of Ethiopian Maccabees 0, The Book of Enoch (Charles) 110, The Book of the Saints of the Ethiopian church 13 (12 months + parent). 2) Fe…

Ancient Christian Writers: The Works of the Fathers in Translation

Paulist Press / The Newman Press (New York-Mahwah, NJ); earlier volumes published by The Newman Press, Westminster MD · English · 78 volumes · 102 works

78 numbered volumes, 1946-2024, all English translations with introduction and annotation; the source languages are Greek and Latin, plus one volume (ACW 16, Irenaeus' Proof of the Apostolic Preaching) translated from Classical Armenian. The series was founded by Johannes Quasten and Joseph C. Plumpe and is now an imprint of Paulist Pres…

Gaps in this series: Clement of Alexandria: Protrepticus, Paedagogus, Stromateis (no ACW volume); Basil of Caesarea: On the Holy Spirit, Hexaemeron, Letters; Gregory of Nazianzus: Theological Orations; Cyril of Alexandria: commentaries and Christological letters; Athanasius: Contra Arianos, De incarnatione (only the Life of Antony is tran…

Ante-Nicene Fathers and Nicene and Post-Nicene Fathers

Christian Literature Company (New York/Buffalo) and T. & T. Clark (Edinburgh); ANF ed. Alexander Roberts & James Donaldson (American edition A. Cleveland Coxe); NPNF Series I ed. Philip Schaff, Series II ed. Philip Schaff & Henry Wace · English · 38 volumes · 126 works

38 numbered volumes: ANF 10 (Ante-Nicene Fathers, American edition 1885-1897), NPNF Series I 14 (1886-1889), NPNF Series II 14 (1890-1900). Volume ANF X is a bibliography/index with no primary texts, so 37 volumes carry translations. The English text is translated from Greek, Latin and (in small part) Syriac originals; the series itself…

Gaps in this series: Coptic: Shenoute of Atripe — the entire corpus (Canons, Discourses) is absent; Coptic: Besa, Life of Shenoute, and Coptic hagiography generally — absent; Coptic: Apophthegmata Patrum in the Coptic/alphabetical collection — absent (only Latin/Greek-derived material appears indirectly); Coptic: Coptic anaphoras and the…

E. A. Wallis Budge's Coptic and Ethiopic translations

Various: British Museum Trustees (5-vol. 'Coptic Texts', 1910-1915); Cambridge University Press; Luzac & Co.; The Medici Society; Oxford University Press; Kegan Paul, Trench, Trubner; privately printed for Lady Meux (W. Griggs) · English, Coptic (Sahidic, 'dialect of Upper Egypt'), Ethiopic (Ge'ez), Greek, Arabic · 20 volumes · 126 works

Verified against the printed volumes themselves, not against secondary bibliographies. The five Coptic volumes are a single series: the front matter of vol. V carries the series title page 'COPTIC TEXTS / EDITED WITH INTRODUCTIONS AND ENGLISH TRANSLATIONS / By E. A. WALLIS BUDGE ... Volume I. Coptic Homilies in the Dialect of Upper Egypt…

Gaps in this series: Shenoute of Atripe: Canons and Discourses (Coptic) - no Budge translation; Coptic liturgy: the Anaphoras of Saint Basil, Saint Gregory and Saint Cyril - not translated by Budge; Coptic Gospels and Psalter (Sahidic New Testament) - Horner's domain, absent here; only Deuteronomy, Jonah, Acts and Revelation appear; Nag H…

Coptic Church and academic English translations outside CSCO

Multiple: British Museum (Coptic Texts, ed. Budge); Oxford: Clarendon Press (Horner, Evetts); D. Nutt (Malan); Brill (Nag Hammadi Studies / Coptic Gnostic Library, History of Christian-Muslim Relations); Cistercian Publications (Cistercian Studies); Cambridge University Press; Oxford University Press; Aris & Phillips; Fortress Press; Coptic Orthodox Church publishers · English · 40 volumes · 139 works

Scope of this corpus: English-language translations of Coptic Christian (and Coptic-preserved Gnostic) literature produced OUTSIDE the Corpus Scriptorum Christianorum Orientalium (CSCO), which itself carries the modern critical Francophone/German/Latin editions and translations of most Coptic texts. This is a corpus, not one publisher's…

Gaps in this series: Shenoute of Atripe: the bulk of his corpus (hundreds of unpublished/unedited works; only ~20 discourses and the Canons are in English); Besa of Athripe: his letters, sermons and other writings (only the Life of Shenoute is translated); The Coptic Synaxarium (the liturgical book of saints' commemorations and lives) - n…

Corpus Scriptorum Christianorum Orientalium

Peeters Publishers, Leuven (series founded 1903 by the Universite catholique de Louvain and the Catholic University of America) · English, French, German, Latin, Italian, Syriac, Coptic, Arabic, Ethiopic, Armenian, Georgian · 713 volumes · 285 works

Verified against the live Peeters catalogue (publisher and sole distributor of CSCO) on 2026-11-13: the series listing at peeters-leuven.be returns 1,422 records - 705 carry a bare main-series number (1-711 and 713), 7 more are combined/asterisked main-series records (47-48-49, 62*, 62**, 63*, 63**, 99-100), and the remaining 710 carry a…

Gaps in this series: 1 Enoch (Ethiopic) - no CSCO edition or translation (0 hits for 'Enoch'/'Henoch' among all 712 records); Peshitta (Syriac Old Testament) - no critical edition in CSCO (that is the Leiden Peshitta Institute); Ethiopic Bible as a whole - CSCO has only John (617/618), Revelation (281/282) and Jubilees (510/511); Apophthe…

ethiopic-english: English translations of Ethiopic literature (Kebra Nagast, Fetha Nagast, PEMM, chronicles) - composite corpus of English translations of Ethiopic (Ge'ez) Christian literature, 1834-2023

Composite / no single publisher: Oriental Translation Fund (1834); Oberlin (1888); Clarendon Press (1896, 1912, 1965, 1978); A. & C. Black (1902); Lady Meux private press (1900, 1906); Young Churchman, Milwaukee (1915); SPCK (1920, 1928); Medici Society (1922, 1923); Cambridge University Press (1928); Methuen (1928); Luzac (1929); Egyptian Book Press, Cairo (1959); Oxford University Press (1932, 1967); Haile Sellassie I University Faculty of Law (1968); Peeters / CSCO (1989); Red Sea Press (1996); Fortress Press (2004); Brill (2015); Princeton University PEMM (2023) · English · 25 volumes · 16 works

COMPOSITE CORPUS, NOT A PUBLISHER SERIES. There is no single 'Ethiopic-English' monograph series published by one house; this entry maps the class of English-language book/corpus translations of Ethiopic (Ge'ez) Christian literature, across 1834-2023. Verification was done from live sources (2026-11-13 UTC): Internet Archive full-text sc…

Gaps in this series: Ethiopic Octateuch (Genesis, Exodus, Leviticus, Numbers, Deuteronomy, Joshua, Judges, Ruth) - no English translation from Ge'ez; Ethiopic Psalter (Mazmura Dawit) and its appended biblical canticles - no English translation from Ge'ez; Ethiopic Ezekiel, Daniel and Isaiah - no English translation from Ge'ez; 1, 2 and 3…

The Fathers of the Church: A New Translation (FOTC)

The Catholic University of America Press (Washington, D.C.) · English · 152 volumes · 338 works

Scope of the count: volumes 1-152 of the main series (1947-2026), all verified as real numbered volumes. The parallel sub-series "Fathers of the Church: Mediaeval Continuations" (at least 26 volumes, e.g. vol. 25 William of Auvergne, On Laws, vol. 26 Albert the Great, On the Sacraments in the publisher's Fall/Winter 2026 catalogue) and "…

Gaps in this series: Shenoute of Atripe (Coptic: Canons and Discourses) - not translated in this series; Besa (Coptic: Life of Shenoute) - not in this series; Pachomian Koinonia (Coptic: Rules, Letters, Lives of Pachomius) - not in this series; Letters of Antony (Coptic; only the Greek Life of Antony appears, in vol. 15); Apophthegmata Pa…

Popular Patristics Series

St Vladimir's Seminary Press (SVS Press), Yonkers, NY · English, Greek (facing-page original in PPS 44A, 22, 62, 57, 65; source languages of the translated texts are Greek, Latin, Syriac, Coptic and Ge'ez) · 68 volumes · 263 works

The series is original English translation, not a critical-text series: Greek is the source for most volumes, with Latin (Tertullian, Cyprian, Gregory the Great, and the Latin sermons in PPS 4), Syriac (Ephrem, Jacob of Serugh, Isaac of Nineveh, the Testament of the Lord), Coptic (only PPS 27 and 28) and Ge'ez/Ethiopic (only PPS 39). Gre…

Gaps in this series: Origen: Against Celsus, On First Principles, Homilies on Genesis/Exodus/Leviticus (only On Prayer in PPS 29 and one Matthew fragment in PPS 48); Clement of Alexandria: Protrepticus, Paedagogus, Stromateis (absent); Justin Martyr: First and Second Apology, Dialogue with Trypho (absent); Irenaeus of Lyons: Against Heres…